THESIS DOCTOR OF PHILOSOPHY KORMAN Bengt 2020

MODELLING OF GAS EXCHANGE
IN THE LUNG DURING NITROUS

OXIDE ANAESTHESIA
Dr Bengt Korman OAM

BSc (Melb), MBBS (Melb), MD (Melb), FFARACS, FANZCA
Submitted in fulfilment of the requirements for the degree of
Doctor of Philosophy
Medical School, Faculty of Health and Medical Sciences
Dept. of Mathematics and Statistics, Faculty of Engineering and Mathematical Sciences
The University of Western Australia 2020

SUPERVISORS
Dr Barry Marshall AC, 2005 Nobel Laureate in Medicine & Physiology
Dr Stephan Schug, Professor of Anaesthesia and Pain Medicine, UWA
Dr Snezhana Abarzhi, Professor of Applied Mathematics, UWA
Dr Philip Peyton, Professor of Anaesthesia & Pain Medicine, University of Melbourne
Last revised 21 September 2020
ii Supervisors
KEYWORDS
Anaesthetic uptake, concentration effect, mathematical modelling, second gas
effect, ventilation-perfusion mismatch
Keywords iii
iv Keywords
ABSTRACT
There are over 4 million alveoli in the adult lung. They are vital to human
life. Uptake of oxygen and elimination of carbon dioxide takes place in these key
structural elements. Gas exchange occurs between alveolar gas and pulmonary
capillary blood and is controlled by the gas/blood flow-ratio. This ratio, termed the
ventilation-perfusion ratio, is normally about 0.8 for the lung as a whole but varies
from alveolus to alveolus, with a spread of values throughout the lung. Past research
indicates that gas exchange will be most efficient if this ratio is the same in each
alveolus and equal to the value for the lung as a whole. A lung with this property is
termed a “homogeneous lung”. Thus, a fundamental principle of respiratory
physiology is that gas exchange can never be more efficient than in a homogeneous
lung.
In 2001, using computer modelling, it was predicted that increased
ventilation-perfusion mismatch could improve gas uptake during nitrous oxide
anaesthesia. The prediction seemed to be confirmed by subsequent experimental
investigations. These findings were ignored by many and disputed by at least one
expert in the field who preferred to look for alternative explanations. These reactions
are not surprising, given that the claims appear to violate the fundamental principle
enunciated above. Therefore, the hypothesis for this research was:
Increased ventilation-perfusion mismatch per se cannot improve gas exchange.
Abstract v
Gas exchange during anaesthesia was modelled by adapting techniques of
formal mathematical analysis developed in the Departments of Medicine and
Mathematics at The University of California, San Diego in the 1970s. Their
mathematical results were further refined for use in individual scenarios. Solutions
were obtained using MATLAB, a programming language with powerful problem-
solving capabilities. The aims of this research fell into three parts as follows:
PART I
To independently replicate the theoretical basis for this strange new
phenomenon. This was necessary to ensure that the previous computer
simulations did not contain undetected errors which might have led to false
predictions.
PART II
To investigate further the effect of nitrous oxide anaesthesia on other gases
present using partial pressure-solubility diagrams. This was necessary because
the previous work on this subject did not take into account the existence of the
new phenomenon.
PART III
To find a plausible explanation of how ventilation-perfusion mismatch can act
to improve gas transfer without violating the fundamental principle of gas
exchange in the lung.
Clearly, parts two and three would only be relevant if the 2001 findings were
confirmed in part one.
vi Abstract
The 2001 computer predictions were independently confirmed. This research
showed that the Minimal Alveolar Concentration or MAC reading does not
accurately reflect the depth of anaesthesia during nitrous oxide anaesthesia when
combined with modern volatile anaesthetics. MAC is used to avoid awareness during
anaesthesia and was the only method used widely for this purpose prior to the
introduction of electronic depth-of-anaesthesia monitors. I also showed that many
results of the formal mathematical analysis needed revision. One unexpected finding
was the possibility of gas uptake seemingly taking place against a partial pressure
gradient.
To explain these findings, I appealed to a previously unrecognized property
of gas exchange, namely that it can be represented as the sum of two steps - step one:
gas exchange at constant volume, followed by step two: gas exchange on volume
correction. The strange findings referred to above, always accompanied the
contraction in volume which occurs during nitrous oxide anaesthesia. Thus, they
were shown to be a product of step two. Accordingly, their magnitude depended on
the amount of gas presented to step two after uptake in step one. By improving
retention at the end of step one, an increase in ventilation-perfusion mismatch exerts
its counterintuitive effect on gas uptake in a way that is perfectly compatible with
classical teaching.
The mysterious phenomena first described in 2001 challenged a fundamental
principle and created a gap in our understanding of gas exchange in the lung. This
gap has now been filled in a way that does not violate the underlying principle. The
research has uncovered several hitherto unrecognized properties of gas exchange and
Abstract vii
led to the discovery of a clinically important feature of nitrous oxide anaesthesia
which can result in patients being too deeply anaesthetized during surgery.
viii Abstract
TABLE OF CONTENTS
: Introduction ........................................................................................... 1
Background .....................................................................................................................1
Context............................................................................................................................3
Purposes ..........................................................................................................................3
Significance, Scope and Definitions ...............................................................................4
Thesis Outline.................................................................................................................6
: Background ............................................................................................ 9
Kety’s theory ..................................................................................................................9
Anaesthetic uptake in peripheral tissues .......................................................................10
Eger’s mathematical model ..........................................................................................11
The second gas effect....................................................................................................13
An additional explanation for the second gas effect .....................................................14
The Eger-Stoelting diagram ..........................................................................................16
The extra-inspired ventilation .......................................................................................18
A matter of relativity ....................................................................................................19
Respiratory pattern during N2O anaesthesia .................................................................19
Respiratory pattern and the Eger-Stoelting diagram.....................................................22
Should Eger have known? ............................................................................................24
A challenge to the validity of the second gas effect .....................................................27
Ventilation-perfusion mismatch appears as a factor .....................................................30
The Severinghaus objection ..........................................................................................31
Summary and Implications ...........................................................................................33
: The Log-Gaussian Distribution .......................................................... 39
Introduction ..................................................................................................................39
Choice of the log normal distribution ...........................................................................40
Correct orientation of graphs ........................................................................................40
The meaning of σ ........................................................................................................40
Measurements of σ in adult subjects ...........................................................................41
The compartmental approach .......................................................................................41
V and Q as explicit functions of log(V/Q)....................................................................43
v and q as explicit functions of (1/σ)log(v/q) ...............................................................46
The case when σ = 0 ....................................................................................................49
Table of Contents ix
Interpretation of gas exchange using Figure 3.4 .......................................................... 49
Interchangeability of x-variables .................................................................................. 50
Lebesgue-Stieltjes integration ...................................................................................... 51
Summary ...................................................................................................................... 53
: Steady-state Equations ........................................................................ 55
The non steady-state model.......................................................................................... 55
The steady-state model................................................................................................. 57
A suitable model to study the second gas effect (SGE) ............................................... 59
Summary ...................................................................................................................... 63
: Persistent Second Gas Effects............................................................. 65
The washin ratio ........................................................................................................... 66
The augmentation ratio ................................................................................................ 67
Simulating persistent second gas effects ...................................................................... 72
Advantages of the mathematical model ....................................................................... 75
Relationship to experimental work .............................................................................. 75
Clinical implications .................................................................................................... 77
Summary and conclusion ............................................................................................. 78
: Partial pressure-solubility diagrams .................................................. 79
The race for MIGET .................................................................................................... 79
Equations...................................................................................................................... 81
The partial pressure-solubility diagram ....................................................................... 86
The augmentation-solubility relationship .................................................................... 92
Alveolar-arterial partial pressure differences ............................................................... 94
The uptake-solubility relationship................................................................................ 98
Partial pressure reversal during gas uptake .................................................................. 99
Alveolar-arterial partial pressure gradient for N2 ....................................................... 102
Conclusion ................................................................................................................. 109
: Elucidating relevant roles ................................................................. 111
Steady-state gas exchange as a two step process ....................................................... 111
Equations for uptake of first gas & diluent gas .......................................................... 116
Equation for inspired ventilation................................................................................ 116
Equation for uptake of second gas ............................................................................. 117
Uptake sequence with a first and second gas ............................................................. 118
Derivation of distributions for diagrams .................................................................... 118
Distributions of inflow, outflow and uptake .............................................................. 122
Second gas uptake as a two-step process ................................................................... 125
Role of the solubility of the second gas ..................................................................... 128
x Table of Contents
Role of V/Q mismatch in relation to the second gas ..................................................129
Effect of V/Q mismatch in relation to the first gas .....................................................130
A third-order contribution to SG uptake .....................................................................132
Location of uptake in the V/Q spectrum.....................................................................133
Effect of ventilation-dependent FG on SG uptake......................................................136
The concentration effect .............................................................................................138
Absorption of gas in airways rather than alveoli ........................................................141
An alternative definition of the Augmentation Ratio .................................................142
Conclusion ..................................................................................................................143
: Influence of Respiratory Pattern...................................................... 145
Introduction ................................................................................................................145
Augmentation Ratios for constant inflow ...................................................................145
The two-step model applied to constant inflow ..........................................................148
Conclusion ..................................................................................................................155
: Conclusions ........................................................................................ 157
Appendix A: Derivation of distribution functions for V and Q ......................... 165
Appendix B: Another version of Kety’s Eqs........................................................ 169
Appendix C: Sample MATLAB routines............................................................. 179
Appendix C.1: ROUTINES USED TO DRAW Figures 6.1, 6.2 and 6.3 .............................179
Appendix C.2 ROUTINES USED TO DRAW Figure 7.4 & Figure 7.5..............................190
Appendix C.3 ROUTINES USED TO DRAW Figure 7.7 ...................................................195
Appendix C.4 ROUTINES USED TO DRAW Figure 8.2 ...................................................198
Bibliography ........................................................................................................... 205
Table of Contents xi
xii Table of Contents
LIST OF FIGURES
Figure 2.1. The concentration effect. ......................................................................... 13

Figure 2.2. The second gas effect. ............................................................................. 14
Figure 2.3. The additional explanation for the second gas effect. ............................. 15
Figure 2.4. The Eger-Stoelting diagram. ................................................................... 16
Figure 2.5. Correctly equilibrated version of the Eger-Stoelting diagram. ............... 18
Figure 2.6. Constant inflow. ...................................................................................... 20
Figure 2.7. Constant outflow. .................................................................................... 21
Figure 2.8. Constant inflow and constant outflow using Eger’s diagram. ................ 23
Figure 2.9. The Forman & Benkwitz version of Eger’s diagram. ............................. 26
Figure 2.10. Letter to Anesthesia & Analgesia by Mapleson & Korman.................. 28
Figure 2.11. Correct version of the Eger-Stoelting diagram. .................................... 33
Figure 3.1. Distributions of ventilation-perfusion ratios. .......................................... 42
Figure 3.2. Ventilation and blood flow as functions of a new variable. .................... 45
Figure 3.3. Distributions of fractional ventilation (v) and blood flow (q). ................ 47
Figure 3.4. Distributions of fractional ventilation (v) and blood flow (-q)................ 48
Figure 3.5. Partitioning the x-axis using a compartmental approach. ....................... 51
Figure 4.1. Washin curve for several anaesthetic gases. ........................................... 58
Figure 5.1. Augmentation ratios for desflurane, isoflurane and diethyl ether. .......... 73
Figure 6.1. Standard partial pressure-solubility diagrams. ........................................ 86
Figure 6.2. Partial pressure-solubility diagrams corrected for volume change. ........ 91
Figure 6.3. Augmentation-solubility diagrams. ......................................................... 93
Figure 6.4. More partial pressure-solubility diagrams............................................... 94
Figure 6.5. Alveolar-arterial partial pressure difference for several gases. ............... 95
Figure 6.6. Uptake-solubility diagrams. .................................................................... 97
Figure 6.7. Alveolar-arterial partial pressure reversal. ............................................ 101
Figure 6.8. A-a partial pressure difference for nitrogen when R < 1. ..................... 105
Figure 6.9. A-a partial pressure difference for nitrogen when R > 1. ..................... 107
Figure 7.1. Distribution of alveolar ventilation and individual gas uptake. ............ 123
Figure 7.2. Gas uptake as a two-step process. ......................................................... 126
Figure 7.3. Uptake profile of SF6, sevoflurane and acetone. ................................... 127
Figure 7.4. Second gas uptake profiles with σ fixed at 2. ....................................... 129
List of Figures xiii

Figure 7.5. Second gas uptake profiles as a function of V Q mismatch. ................ 131
Figure 7.6. 3-dimensional representation of second gas uptake. ............................. 133
Figure 7.7. Ventilation and perfusion dependent uptake of inert gases. .................. 134
Figure 7.8. Effect of volume contraction with N2O and diethyl ether. .................... 137
Figure 7.9. Two-step analysis of the concentration effect. ...................................... 138
Figure 7.10. Influence of solubility on step 2 of concentration effect. .................... 140
Figure 8.1. Persistent second gas effects with constant inflow. .............................. 146
Figure 8.2. Comparison of constant outflow and constant inflow. .......................... 153
Figure 8.3. Augmentation ratios for constant inflow and constant outflow. ........... 154
Many of the figures in this thesis have been prepared with the help of the Department
of Medical Illustration, Royal Perth Hospital. In spite of their expert efforts, some of
the figures exhibit discontinuities and/or pixilation when viewed on a computer
screen. These effects may be reduced or eliminated by trying a different page
magnification e.g. 75% or 100%.
xiv List of Figures

LIST OF TABLES
Table 5.1. The nitrous oxide Washin Ratio…………………………………..66

Table 7.1. Gas exchange in a lung with 4 compartments……………………115
Table 7.2. List of vectors used in obtaining distributions of gas uptake….…120
Table 7.3. Factors affecting gas uptake in each step of a mechanistic model of
gas exchange……………………………………………………….143
Table B.1. Tissue volume, blood pool and blood flow for various
compartments………………………………………………………171
Table B.2. Tissue/gas partition coefficients for various agents in several
tissues………………………………………………...………………172
List of Tables xv
xvi
LIST OF ABBREVIATIONS
General variables:
V Gas volume in general

V Gas volume per minute
P Pressure, mmHg
F Fractional concentration of gas in dry phase
Q Volume of blood flow, l/min.
C Concentration in blood phase
λ Partition coefficient between two phases at equilibrium (usually blood/gas)
σ Degree of mismatch between ventilation and blood flow
AR Augmentation ratio
R(λ) Ratio of gas exchange rates
Symbols for gas phase:
I Inspired gas
E Expired gas
A Alveolar gas
B Barometric
Symbols for blood:
a Arterial
v Mixed venous
Common abbreviations:
FG First gas
SG Second gas
CE Concentration effect
SGE Second gas effect
Other context-specific abbreviations are defined as the need arises.
List of Abbreviations xvii

ASSUMPTIONS MADE IN MODELLING
The models presented in this thesis contain many assumptions. Most of
these were necessary to facilitate the mathematical treatments. They include the
following:
1. Oxygen and carbon dioxide are excluded from the models because
of their complicated interactions with haemoglobin
2. All other gases obey Henry’s Law
3. Unless stated otherwise, steady-state conditions apply
4. The functional residual capacity is assumed to remain constant
during gas exchange
5. Absorption atelectasis and hypoxic vasoconstriction are not
considered.
These assumptions are discussed in further detail in relevant sections of the
thesis.
xviii Assumptions made in Modelling

PUBLICATIONS
Details of the work:

Korman B, Dash RK, and Peyton PJ. Can mathematical modeling explain the measured magnitude
of the second gas effect? Anesthesiology 128: 1075 - 1083, 2018.
Location in thesis: Chapter 5
Student contribution to work:
Conception, mathematical modelling, computer simulation, assessment of results, preparation of
illustrations, preparation of manuscript, corresponding author.
Co-author signature Co-author signature
Philip J Peyton Ranjan K Dash
Date: Date: 10/14/2019

Korman B, Dash RK, and Peyton PJ. Effect of net gas volume changes on alveolar and arterial gas
partial pressures in the presence of ventilation-perfusion mismatch. J Appl Physiol 126: 558-568,
2019.
Date: Date: 10/14/2019

Korman B, Dash RK, and Peyton PJ. Elucidating the roles of solubility and ventilation-perfusion
mismatch in the second gas effect using a two-step model of gas exchange. J Appl Physiol, 128:
1587-1593, 2020. https://doi.org/10.1152/japplphysio.00049.2020
Date: 17/03/2020 Date: 03/17/2020
Publications xix
These publications are appended to the end of this thesis. Permission to include
the first manuscript was granted by ANESTHESIOLOGY and Wolters Kluwer Health
Inc. (publishers of the journal and copyright owners). An editorial on the article
written by Professor Ross Kennedy may be viewed at the ANESTHESIOLOGY
website (Kennedy RR. A Second Look at the Second Gas Effect. Anesthesiology
128: 1053-1054, 2018 (https://anesthesiology.pubs.asahq.org/article.aspx?articleid=2677923).
The American Physiological Society permits use of published material in theses.
A review of concentration and second gas effects, based on this thesis, is
currently being prepared.
xx Publications
STATEMENT OF ORIGINAL AUTHORSHIP
I, Bengt Korman, certify that:
The work contained in this thesis is my work, has been completed during
the course of my degree and does not breach any ethical rules with regard to the
conduct of the research. It has not been previously submitted to meet
requirements for an award at this, or any other higher education institution.
To the best of my knowledge and belief, the thesis contains no material

previously published or written by another person except where due reference is
made in the text and, where relevant, in the Authorship Declaration that follows.
This thesis does not violate or infringe any copyright, trademark, patent,
or other rights whatsoever of any person.
In future, no part of this thesis will be used in a submission in my name,

for any other degree or diploma in any university or other tertiary institution
without the prior approval of The University of Western Australia and where
applicable, any partner institution responsible for the joint award of this degree.
This thesis contains published work and work prepared for publication
which has been co-authored.
Signature: Ben Korman
Date: 22 September 2019
Statement of Original Authorship xxi

MENTORS
Ian M Ritchie William W Mapleson
Alexander P Robertson
xxii Mentors
ACKNOWLEDGEMENTS
As soon as I walked into his office in the School of Chemistry at UWA in
1976, Professor Ian Mackay Ritchie (73) welcomed me warmly. As an old
student from his days in the Department of Chemistry at The University of
Melbourne, I knew that he possessed those attributes I admired most – a sharp
mind, humility, kindness, generosity, and a keen sense of humour. Ian had
supreme confidence in my ability and encouraged me to investigate those strange
mysteries that seemed to pop up repeatedly during my training as an anaesthetist.
Ian’s boundless energy and enthusiasm inspired all those who came into contact
with him. A true friend and inspirational mentor.
Ian helped me establish my credentials with Professor William Wolseley
Mapleson, Professor of the Physics of Anaesthesia at the University of Wales
(35), one of the great pioneers of anaesthesia as a science. Bill helped me clarify
abstract ideas that led to the publication of two papers in joint name in the British
Journal of Anaesthesia.
Ian also introduced me to Professor Alexander Proban Robertson,
Foundation Professor of Mathematics at Murdoch University (6), whose talent
for simplifying the most complicated mathematical expressions was a delight to
witness. He and Ian Ritchie were both graduates of Cambridge University.
Sadly, all three of these great teachers have now passed on. This thesis
is a tribute to their skills, their patience and their dedication to science.
Acknowledgements xxiii
Other people influenced my decision to undertake this research and
assisted me in conducting it. Dr Philip Peyton, Professor of Anaesthesia and Pain
Medicine at The University of Melbourne, my alma mater, suggested that I enrol
for the degree of PhD at the university. This proved impossible because I was
still working in full-time private practice in Perth and could therefore not spend
the requisite time in Melbourne under his supervision. Nevertheless, he offered
to be an External Supervisor for a PhD degree undertaken at The University of
Western Australia. In this role, he carefully checked every aspect of my
theoretical work, drawing possible errors to my attention and preventing me
from making serious mistakes when submitting manuscripts to Reviewers of the
most prestigious journals in the specialties of Anaesthesia and Respiratory
Physiology.
Dr Stephan Schug, Professor of Anaesthesia and Pain Medicine at The
University of Western Australia kindly agreed to act as Co-ordinating
Supervisor. Even though the work was somewhat outside his ambit, he
enthusiastically supported me and praised each new development in the theory.
I could not have wished for a more supportive person in this role.
Drs Snezhana Abarzhi and Michael Small, each a Professor in the
Department of Applied Mathematics at UWA, made contributions. Moreover, I
was extremely fortunate to have formed a working relationship with Dr Ranjan
K. Dash, Professor of Bioengineering and Physiology at the Medical College of
Wisconsin and Marquette University. He introduced me to MATLAB, the
computer language that helps eliminate the drudgery of solving numerical
problems. He generously assisted in debugging faulty code, so smoothing my
path through difficult programming terrain.
xxiv Acknowledgements
Dr Barry Marshall, Professor of Medicine at UWA, Nobel Laureate in
Medicine in 2005, graciously agreed to lend his name as Local Supervisor with
a recent PhD candidate, an obligatory requirement for those undertaking the
examination. This requirement was not satisfied by any of the other Local
Supervisors.
Finally, I would like to acknowledge my wife Helen without whose
patience and understanding this project would not have been possible.
This research was supported by an Australian Government Research
Training Program (RTP) Scholarship.
Acknowledgements xxv
xxvi Acknowledgements
: Introduction
BACKGROUND
In 2014 I felt it was my turn to present a talk at the regular Thursday
morning meetings held in the Department of Anaesthesia and Pain Medicine at
the Royal Perth Hospital. I chose to talk about a subject that had formed part of
my MD thesis for The University of Melbourne in 1996, a diagram purporting
to explain certain aspects of gas uptake during anaesthesia. These were the
concentration effect and the second gas effect.
The diagram by Edmond Eger, one of the pioneers in the field of
pharmacokinetics of anaesthetic gases, first appeared in 1963 (19) and had
subsequently appeared in numerous anaesthetic textbooks. I was at first
entranced by the diagram, its apparent simplicity, its ability to convey a
complicated message so clearly. But over time, I began to feel there was
something wrong with the diagram. My suspicions were confirmed when Ian
Ritchie, then Associate Professor of Chemistry at UWA pointed out a flaw. If
Eger’s diagram was wrong, then what was the correct diagram? It took me many
years to sort out the puzzle. After publishing a joint paper on the subject with
Professor Mapleson in 1997 (48), Eger’s diagram virtually disappeared from all
textbooks except those written in the United States, his home country.
In preparing for my proposed talk to the Department, I undertook a review
of material published on the subject since 1997. To my surprise, I found that
another Australian anaesthetist, Philip Peyton had published several papers on
the same subject in the Journal of Applied Physiology, in which he claimed that
Chapter 1: Introduction 1
mismatching of ventilation and blood flow can lead to improved gas uptake (68-
70). These theoretical papers appeared to be supported by experimental evidence
from Peyton himself and another group based in Belgium (34, 67, 71). Being
busy in a full-time private practice I had not kept up with the literature and was
taken completely by surprise by these claims.
Classical teaching requires that if gas exchange is to proceed most
efficiently, ventilation, the process of gas delivery to the gas-exchanging units
in the lung, should be perfectly matched to the blood flowing through the lung
(26). If the ventilation is for example, 4 l/min and the pulmonary blood flow is,
say, 5 l/min, this entails the ratio of gas flow to blood flow to be 4/5 or 0.8 in
each gas-exchanging unit throughout the lung. Here was someone claiming the
opposite can be true, that a worsening in the matching of ventilation and blood
flow throughout the lung can actually improve gas exchange. I found this quite
astounding. My amazement was compounded by the fact that a journal as
prestigious as the Journal of Applied Physiology could have allowed such heresy
to appear in its pages. I knew only too well how difficult it is to have work appear
in the journal having had submissions rejected on at least two occasions.
I believed that Peyton may have erred either in his original assumptions or
in his computer simulations. I therefore set out with the aim of replicating his
steps and locating his mistake. This would require an algorithm to deal with the
non-linear interaction of oxygen and carbon dioxide with haemoglobin, as
Peyton had included these physiological gases in his model. The algorithms he
used were those of Kelman, published in the 1960s (38, 39). Surely a more up-
to-date algorithm was now available. A quick internet search yielded a paper by
Dash and Bassingthwaighte (13). At first, I struggled with the chemistry and
2 Chapter 1: Introduction
maths in the paper but after reading it several times, I realized that it was quite
straightforward. Moreover, I found what I thought was a short-cut which could
speed up calculations considerably. James Bassingthwaighte, a Professor in the
Department of Bioengineering, Biomathematics and Radiology at the University
of Washington in Seattle, was unsure but Ranjan K. Dash, then Associate
Professor of Bioengineering at The University of Wisconsin immediately
agreed. The three of us subsequently co-authored a paper in the European Journal
of Applied Physiology (14). It included my short-cut as well as other
improvements by Ranjan Dash.
So began a productive relationship between the two of us. Ranjan advised
me to change my programming language from Visual Basic to MATLAB, a
widely used scientific programming language well-suited to solving the
complicated expressions that kept arising in my investigations. It very quickly
became apparent that Peyton’s findings could not be rejected. When I discussed
this with him, Phil suggested I enrol for the degree of Doctor of Philosophy with
him as a Supervisor. Together, we could investigate the phenomenon further.
This thesis is the result of that collaboration.
CONTEXT
The major focus of this study is the proposition that gas uptake can be
improved by ventilation-perfusion mismatch, a notion that appears to completely
contradict one of the fundamental principles of respiratory physiology.
PURPOSES
The purpose and specific aims are as follows:
1. To develop a suitable model for investigating the concentration & second
gas effects (CE & SGE)
2. To use this model to study the effects on the CE & SGE of varying solubility
and ventilation-perfusion mismatch
3. To find clinically relevant predictions using the model
4. To explain how ventilation-perfusion mismatch acts to improve gas uptake.
SIGNIFICANCE, SCOPE AND DEFINITIONS
The widespread clinical use of nitrous oxide led to the initial discovery of
the concentration and second gas effects. However, the use of this agent has
diminished over recent years because of
1. Concerns regarding its safety
2. Concerns regarding its environmental impact
3. The increasing popularity of total intravenous anaesthesia.
Nitrous oxide has traditionally been regarded as a very safe agent, having
been used in anaesthesia for over 150 years. Minor issues were considered to be
the possible depletion of vitamin B12 stores associated with long-term exposure
especially in subjects with vitamin B12 deficiency (28), and nausea and vomiting
(15, 32, 87). The ENIGMA trial, published in 2007 (62), led to the conclusion
that the administration of nitrous oxide was associated with increased risk of
perioperative wound infection and long-term risk of myocardial infarction (51).
Although these fears were not borne out by the larger and more powerful
ENIGMA 2 trial (63), many anaesthetists had by then abandoned it completely
and new trainees in anaesthesia were taught to avoid it wherever possible. The
environmental impact is clearly of concern given that the gas is said to have an
effect on the ozone layer, and a contribution to global warming some 300 times
that of carbon dioxide on a weight for weight basis (77). Moreover, the effect is
very long lasting compared with that of many of the carbon-containing
compounds. Nevertheless, a report on the environmental impact of nitrous oxide
in member countries prepared by the European Economic Community in the
1990s regarded the contribution from medical use to be very small (11). The gas
was included in the 2017 WHO List of Essential Medicines and is still available
in most operating theatres in Australia. The combination of nitrous oxide and
sevoflurane remains a popular choice for gaseous induction in paediatrics, and
nitrous oxide is still used widely in operating theatres, labour wards and dental
practices.
Even if nitrous oxide were to disappear completely from clinical use, it
is my view that the awkward facts discovered by Peyton cannot merely be swept
under the carpet. They challenge a basic tenet of respiratory physiology and
demand investigation. We anaesthetists have stumbled onto this “inconvenient
truth” and if the respiratory physiologists are not prepared to investigate its
cause, then surely it falls to the anaesthetic fraternity to accept responsibility and
come up with a plausible explanation.
The approach used in this thesis is purely theoretical. Since using gas
chromatography, mass spectrometry and infra-red spectroscopy in the 1980s as
part of the research for my MD thesis, I have not been involved in any
experimental work. It could be argued that each anaesthetic I have administered
constitutes a separate experiment and that the skills involved might easily be
translated into experimental research. Each of us must, however, recognise our
limitations - I am well aware of mine. Experimental work is not my forte. On the
other hand, over recent years my fascination with mathematics has grown.
Having been exposed to the rigorous development of the theory of linear algebra
and of series in second year Pure Mathematics at honours level at the University
of Melbourne, I know just how powerful a tool, mathematical methods can be
when applied appropriately. What seemed dull and uninteresting in 1966 became
important and relevant in tackling the problems that arose repeatedly. In regard
to the matter at hand, I would argue that the important experimental work had
already been done by Peyton, when he confirmed his prediction that the second
gas effect is greater in blood than in the gas phase and that this is due to the
increase in ventilation-perfusion mismatch which accompanies induction of
anaesthesia. In my view, what was required now was not more experimental
work, but a satisfactory explanation of the cause. This could only come from
theory.
THESIS OUTLINE
Chapter 2. This introduces the topic. I describe how a diagram purporting
to explain important aspects of gas uptake was fundamentally flawed but was
blindly accepted and reproduced ad nauseum much to the confusion of the
anaesthetic community. I move on to describe the main focus of my attention,
the proposition that ventilation-perfusion mismatch can somehow improve gas
uptake. The history of this proposition is traced from inception to the time I first
became aware of its existence.
In Chapters 3 and 4 important mathematical concepts are introduced.
These are central to the theoretical investigations that follow.
Chapter 5 addresses the issue of persistent second gas effects, claimed to
exist by several workers in the area. I set out to confirm or deny the existence of
these effects.
In Chapter 6, concepts developed in the preceding chapters are extended
in an attempt to generalize the findings. As a framework, I use the partial
pressure-solubility diagrams developed by Respiratory Physiologists at the
University of California in San Diego over 40 years ago.
In Chapter 7, gas exchange is treated as a two-step process in which gas
volume contraction is confined to step 2. This finally provides a plausible
explanation for the effect of V Q mismatch observed clinically with the low
solubility volatile agents in use today.
In Chapter 8, the two-step analysis is used to compare the effect of
different respiratory patterns on gas uptake in the presence of V Q mismatch.
Chapter 9 summarizes the findings presented in this thesis.
Where the theory relies on certain mathematical results or treatments, these
are usually introduced at the start of the relevant chapter. When the theory begins
with the same assumptions as in previous chapters, the assumptions are often re-
stated so as to avoid the need to move backwards and forwards in the thesis. It
is hoped that this will smooth the reader’s task and prevent unnecessary
interruption in the line of thought.
: Background
In 1951, Kety published a comprehensive review of inert gas exchange at
the lungs and tissues (42). He identified factors which determine the rate of gas
uptake and indicated the direction to be followed in subsequent research.
KETY’S THEORY
Expressed using symbols with which we are familiar today, Kety produced
the following equation for the rate of change with time of the alveolar partial
pressure of an inert gas i.e. a gas which merely dissolves in blood and does not
interact with it chemically (see also discussion in the next section):
dPA 
VL = VA ( PI − PA ) + λQ t ( P v − Pa ) (2.1)
dt
where,
VL = alveolar air volume
VA = alveolar ventilation
PI = partial pressure of the inert gas in the inspired gas mixture
PA = partial pressure of the inert gas in mixed alveolar gas
Q t = pulmonary capillary blood flow
P v = partial pressure of inert gas in mixed venous blood
Pa = partial pressure of inert gas in arterial blood
λ = blood-gas partition coefficient
Chapter 2: Background 9
Although Kety realized that the body tissues could not be treated as a
homogeneous tissue mass, the simplest solution was achieved by making this
assumption with respect to blood flow and solubility, together with the
assumptions that equilibrium between blood and tissue is complete and that the
blood/tissue partition coefficient is 1. Assuming a tissue volume of VT litres, this
gives rise to an equation of the following form for events at the tissues:
dP v
VT= Q t ( Pa − P v ) (2.2)
dt
Since Pa = PA in a homogeneous lung, Eqs. 2.1 and 2.2 may be solved by
standard means to give a solution of the form:
PA =PI (1 − A 1 e − k1t − A2 e − k2t ) (2.3)
where A1 , A2 , k1 , k2 are constants. Importantly, this solution predicts that
PA PI as a function of time, is independent of PI .
ANAESTHETIC UPTAKE IN PERIPHERAL TISSUES
In the period between 1951 and 1963 anaesthetic interest was directed at
refining Kety’s solution to improve the fit to experimental data. Copperman,
cited by Kety, allowed for the division of the body into a number of
compartments (42). The tissue/blood partition coefficient of each tissue
compartment could then be adjusted appropriately, instead of assuming a value
of 1 as Kety had done. The blood flow per unit of tissue volume was the same in
any one compartment but differed from one compartment to another.
The solution of the relevant equations was complicated by the huge
number of calculations required and was initially achieved using an electric
10 Chapter 2: Background
analogue. Although several analogues were presented at a gathering of experts
invited to a famous conference in 1962, there is no doubt in my mind that the
first electric analogue was that of Mapleson (57). His paper was submitted to the
prestigious Journal of Applied Physiology on 4th December 1961 but was not
published until January 1963! At least one of the other participants who
presented an electric analogue at the 1962 conference was also a Reviewer for
the journal. One wonders how this coincidence would be viewed today.
EGER’S MATHEMATICAL MODEL
At the conference in 1962, Eger produced a mathematical model of uptake
and distribution of anaesthetic gases (18, 21). We can summarize the relevant
parts of his model as follows 1:
VL • Cg n −1 + VI • FI − un + FI • un
Cg n = (2.4)
VL
 +u
Cbn −1 • QT
Cbn = n
(2.5)

QT
Cbn
λ= (2.6)
Cg n
where:
Cg n = concentration of anaesthetic agent in gas phase at the end of the nth breath
1
Eger included corrections for temperature, humidification and volume of lung tissue but these are not
critical to the discovery of the concentration effect and resulted in his model becoming unnecessarily
complicated. We do not include them in this thesis as inspired gas is considered to be fully humidified
at 37 deg C and we use the fractional concentration in dry gas as our measure of concentration. Note
that events in VL , the lung volume, are irrelevant in the steady-state model of gas exchange used to
study the concentration and second gas effects in this thesis (see Chapter 4).
Cbn = concentration of anaesthetic agent in blood at the end of the nth breath
VL = volume of air present in the lungs at the end of passive expiration 2
VI = volume of inspired gas mixture delivered during the nth inspiration
un = uptake of anaesthetic during the nth breath
Q = pulmonary blood flow
T = duration of each breath
The numerator of Eq. 2.4 states that the volume of the anaesthetic gas
present in the lung at the end of the nth breath is equal to the volume of
anaesthetic gas in the lung at the beginning of the nth breath plus the volume of
the anaesthetic gas brought in during the breath minus the volume of anaesthetic
gas absorbed by equilibration with blood plus the volume of extra fresh gas
mixture drawn in to replace the volume of anaesthetic gas transferred to blood.
Eq. 2.6 is the form of Henry’s Law commonly used in Anaesthesia and
Respiratory Physiology. A gas which obeys this law is said to be an inert gas.
From his model, Eger predicted that FA FI would rise more rapidly for higher
values of FI . This is shown in Figure 2.1 and was subsequently confirmed
experimentally (19). He named this phenomenon the concentration effect
because it is significant for agents administered at high inspired concentrations,
particularly if the agent is very soluble in blood. In Figure 2.1, the effect is seen
to be greater for the more soluble agent ether, than for nitrous oxide 3.
2
This is the functional residual capacity or FRC.
3
For copyright reasons, this diagram and several others in this thesis are simulated using inputs from
the original publication in a new computer model described in Appendix B.
Figure 2.1. The concentration effect.
Alveolar concentration as a fraction of inspired concentration for inspired concentrations
of 1%, 40%, 75% & 100% N2O and 1%, 40% and 75% ether during washin. From a
model described in Appendix B with inputs based on the original material by Eger (18).
THE SECOND GAS EFFECT
Because it is a weak anaesthetic agent, it is common to supplement nitrous
oxide with low concentrations of a more potent volatile anaesthetic. Using
mixtures of nitrous oxide and halothane, Epstein et al. (24) showed that when
the concentration effect is in operation, it may accelerate the rise of FA FI for a
second agent administered simultaneously (Figure 2.2). Although the inspired
concentration of halothane was the same in both cases, FA FI rose more rapidly
in the presence of 70% nitrous oxide – the second gas effect. This effect was
interpreted by the authors to be the result of the additional respiratory inflow
secondary to the absorption of nitrous oxide at higher concentrations i.e. the term
FI • un in Eq. 2.4 above. The effect was subsequently documented for oxygen
(33) and carbon dioxide (43).
Figure 2.2. The second gas effect.
Simulation of the experiment by Epstein et al. (24) using a model described in Appendix
B, with inputs based on the original investigation in which 0.5% halothane was
administered to dogs anaesthetized with pentobarbital using one of two possible
anaesthetic mixtures. One mixture included 70% nitrous oxide, the other, 10% nitrous
oxide. 4
AN ADDITIONAL EXPLANATION FOR THE SECOND GAS EFFECT
In a subsequent investigation, Stoelting and Eger equilibrated dogs with
low concentrations of ethylene, cyclopropane or halothane in oxygen (84). The
inspired gas composition was then changed abruptly to a mixture containing 70%
nitrous oxide, the equilibrium concentration of the second gas and oxygen. FA
for the second gas was then observed to rise above FI (see Figure 2.3). The
authors argued that an increase in the inspired ventilation could not explain these
4
The original article may be viewed at:
https://anesthesiology.pubs.asahq.org/article.aspx?articleid=1966256
Figure 2.3. The additional explanation for the second gas effect.
Simulation of experiment by Stoelting and Eger (84) using a model described in Appendix
B with inputs based on the original investigation in which dogs were equilibrated with
low concentrations of ethylene (blue, λ = 0.14), cyclopropane (green, λ = 0.415) and
halothane (red, λ = 2.3) in oxygen. At the time indicated as zero, the inspired gas
composition was abruptly changed to a mixture of 70% nitrous oxide containing the
previous concentration of second gas. The balance of the inspired gas mixture consisted
of oxygen. The alveolar concentration of each second gas was then observed to rise above
the inspired concentration. Ventilation was controlled with a volume-limited ventilator. 5
results as it would oppose the rise of FA above FI. They postulated that an
additional factor, a concentrating effect must be involved. They illustrated their
results by using the diagram previously employed by Eger to explain the
concentration effect (19).
5
The original article may be viewed at:
THE EGER-STOELTING DIAGRAM 6
Figure 2.4. The Eger-Stoelting diagram.

Diagram of hypothetical lung used by Eger and Stoelting to explain the concentration and
second gas effects. Reproduced from Stoelting and Eger (84) with permission from copyright
owners Wolters Kluwer Health Inc.
The revised diagram was now used to explain both the concentration
and second gas effects. Since its first appearance in ANESTHESIOLOGY in 1969 it
has been reproduced in numerous anaesthetic textbooks whenever the effects
are discussed. The explanation in the caption given by the authors is as
follows:
“The hypothetical lung initially contains 80% nitrous oxide, 19% oxygen
and 1% second gas. A – the concentrating effect. If half the nitrous oxide is taken
up, the remaining second gas now represents 1.7% of the total gas volume, while
before it represented only 1%. Consequently, the second gas has been
concentrated in a smaller gas volume and its alveolar concentration increases.
B – the increased inspiratory ventilation. This is necessary to
maintain lung volume. The inflowing gas contains the same proportions of
nitrous oxide, oxygen and second gas as the gas originally present. Although
this additional
6
The original article containing this diagram may be viewed at
ventilation increases the nitrous oxide concentration from 66.7% to 72%, it
dilutes the previously-concentrated second gas and diminishes the magnitude of
the second gas effect.”
No doubt Eger believed the diagram to accurately reflect the steps in his
mathematical model. Sadly, as we will now demonstrate, this was not the case.
Notice that as the rule for equilibrating nitrous oxide with blood, Eger is
postulating that half the nitrous is taken up. However, the rule is only applied to
the initial gas. The extra gas brought in to replace the nitrous oxide taken up by
blood is left unequilibrated in the lung. Looking at Eq. 2.4, we can see that the
extra gas is included in the calculation of Cg n which is itself then equilibrated
with blood in Eq. 2.6. Therefore, the diagram and the model are NOT the same.
Moreover, the notion of gas sitting for any length of time in the
lung, unequilibrated with blood is incompatible with basic assumptions of
inert gas exchange, a well-known fact by 1969 (9, 30). Equilibration with
blood leads to the situation shown in Figure 2.5 (44). The starting situation is
shown on the left of the figure. Now however, the extra-inspired ventilation has
been added to the original volume V giving a total volume of 12/3 V .
Nitrous oxide comprises 80% of the total volume or 11/3 V. Following
proper equilibration with blood according to Eger’s formula that half the
nitrous oxide be taken up, 2/3 V of the nitrous oxide disappears into blood
leaving 2
/3 V behind in the gas phase.
Figure 2.5. Correctly equilibrated version of the Eger-Stoelting diagram.
Reproduced in modified form from Korman (44).
THE EXTRA-INSPIRED VENTILATION
From the time he first gave his explanation of the concentration and second
gas effects, Eger routinely invoked the twin mechanisms of the concentrating
effect and the extra inspired ventilation. Presumably because the concentration
effect was greater with the more soluble agent ether than with nitrous oxide 7, he
postulated that the extra-inspired ventilation is more important for more soluble
agents, while based on the findings shown in Figure 2.3, the concentrating effect
is more important for less soluble agents. Thus, the extra-inspired ventilation
became a standard component of discussions of these effects with the clear
message that it is always present.
7
This depends on how one measures the concentration effect. In  7.15, we show that when inspired
concentrations are chosen so as to produce a similar contraction in gas volume, the effect is actually
greater for the less soluble agent, nitrous oxide. This involves using 25% ether in a comparison with
70% nitrous oxide.
A MATTER OF RELATIVITY
The question of whether an extra-inspired ventilation is always a part of
gas exchange was the subject of many conversations between Ian Ritchie and
me. Ritchie pointed out that it depends where the observer is located. If the
observer is located at the alveolar-capillary junction, large volumes of different
gases are constantly flowing in both directions. During induction of anaesthesia
with a gaseous agent, the net movement of anaesthetic gas is into blood. The
observer stationed there will perceive that more anaesthetic gas is drawn down
the airway to replace that taken up by blood. This must always occur at the
interface between gas and blood if anaesthetic uptake is to proceed without
interruption.
However, we anaesthetists are not stationed at the alveolar-capillary
membrane. We are standing beside our patients and observing events from the
outside. This is associated with a different perspective which will now be
considered.
RESPIRATORY PATTERN DURING N2O ANAESTHESIA
Kety (42), Mapleson (57) and Eger (21) all treated the functional residual
capacity (FRC) as remaining constant during anaesthetic uptake. Any changes
in volume due to gas uptake can then only be reflected in such models by
differences between the inspired and expired ventilation. In particular, VIA , the
inspired alveolar ventilation, must exceed VEA , the expired alveolar ventilation.
This gives rise to 2 possible extreme patterns of respiration. We have named
these constant inflow and constant outflow.8
8
These terms were first suggested by Professor Alex Robertson.
2.9.1 CONSTANT INFLOW
In this pattern, the inspired tidal volume is kept constant and the expired
tidal volume allowed to vary, reflecting the gas uptake during the breath.
Figure 2.6. Constant inflow.

Reproduced with permission from Korman and Mapleson (48).
In Figure 2.6, each pair of vertical lines represents the volume of an
inspiration followed by the accompanying expiration. Inspiratory volume is the
same for each breath. Expiratory volume varies, depending on the net gas
volume uptake during that breath. The first pair of lines shows the situation
before the anaesthetic gas is introduced when the difference in volume is due to
oxygen uptake exceeding carbon dioxide output and is so small as to be
imperceptible. Thereafter, the differences between inspiration and expiration
become apparent as large volumes of nitrous oxide are taken up by blood during
induction with 70% N2O. As anaesthetic washin proceeds, FA FI approaches 1
and the expired tidal volume slowly returns to its pre-induction value. The
situation is approximated clinically by a constant volume ventilator which
delivers the same volume with each breath.
2.9.2 CONSTANT OUTFLOW
At the other extreme, we have a constant outflow pattern in which the
expired tidal volume is fixed, and the inspired tidal volume allowed to vary.
Figure 2.7. Constant outflow.

Reproduced with permission from Korman and Mapleson (48)
In Figure 2.7, the second line in each pair represents the constant expired
tidal volume. To maintain the expired tidal volume constant, it is necessary for
the inspired tidal volume to exceed it. The amount by which the inspired tidal
volume exceeds that expired is the volume of nitrous oxide uptake during that
breath. This volume is indicated by the stippled area and may be thought of as
being drawn into the airway to maintain the sum of the FRC and expired tidal
volume at a constant value. We can therefore identify this gas as Eger’s extra-
inspired ventilation. This pattern is approximated clinically by a spontaneously
breathing subject attempting to maintain a constant arterial PCO2 .
2.9.3 EGER’S CONSTANT INFLOW MODEL
It is important to note that while Eger produced equations of the form of
Eqs. 2.4 - 2.6 for the constant outflow case (which he equated to a nonrebreathing
system 9), he also produced the following equation for the constant inflow case:
VL • Cg n −1 + VI • FI − un
Cg n = (2.7)
VL + VI • FI − un
Equations 2.5 and 2.6 were applied as before. On rearrangement, this system of
equations gives rise to a quadratic equation which is solved for un . In this system
also, the higher the inspired concentration, the more rapid the approach to the
final concentration but not as fast as with Eqs. 2.4 - 2.6 (18). Eger attributed the
reduction in rate to the limited inflow associated with the use of a circle system 10.
He commented that the solution “is not difficult, but cumbersome”.
RESPIRATORY PATTERN AND THE EGER-STOELTING DIAGRAM
Once gas in the lung is fully equilibrated with blood, we may draw the
following version of the corrected Eger-Stoelting diagram to illustrate the
difference between the two extreme respiratory patterns (48). In Figure 2.8, we
have again applied the rule that half the nitrous oxide be taken up. Note that the
constant inflow case is associated with a smaller uptake of nitrous oxide than the
equivalent constant outflow case (160 ml compared with 267 ml). This is a
feature of constant inflow and reflects the limitation on inflow referred to by
9
This is incorrect as a non-rebreathing system used with a constant volume ventilator actually has a
constant inflow pattern.
10
Again, this is wrong since a subject breathing spontaneously on a circle system with an adequate
fresh gas flow will have a constant outflow pattern of respiration.
Eger. When we apply the rule that half the nitrous oxide be taken up, the final
concentration is the same with both patterns.
Once the FRC is fixed, the only possible patterns are constant inflow,
constant outflow and all combinations thereof. Since the concentration effect
occurs in both extreme cases, it follows that it must occur in all combinations
thereof. But there is no extra-inspired ventilation with constant inflow. As a
result, the extra-inspired ventilation cannot be credited with causing
Figure 2.8. Constant inflow and constant outflow using Eger’s diagram.
A-B: Inspired tidal volume is kept constant and equal to 400 ml. C-D: Expired tidal
volume is kept constant and equal to 400 ml. BE = Before Equilibration; AE = After
Equilibration. The gradations on each side of the y-axis occur at intervals of 20% of the
inspired tidal volume for each case, that is at intervals of 80 ml for the constant inflow
case and 133.3 ml for the constant outflow case. Reproduced with permission from
Korman & Mapleson (48).
the concentration and second gas effects. Instead we must look for some other
property common to both constant inflow and constant outflow. That property,
evident in all our figures from Figure 2.4 to Figure 2.8, is the shrinkage in volume
and the associated concentrating effect. Thus, we conclude that the concentration
and second gas effects are ALWAYS caused by the concentrating effect that
accompanies net gas volume uptake.
SHOULD EGER HAVE KNOWN?
Eger’s presentation at the 1962 conference chaired by Papper and Kitz was
immediately followed by a commentary given by Herbert Rackow (74). He
referred to previous work published by his group in 1959 in the Journal of
Applied Physiology in which the authors described the effect of different
ventilators during emergence from nitrous oxide anaesthesia (75). In so doing,
he actually presented the constant inflow and constant outflow patterns. He went
on to comment that he thought this was part of Eger’s concentration effect.
Strangely, neither Eger nor Rackow seems to have made the connection.
For year after year, Eger went on explaining the concentration effect repeatedly
using his diagram and always invoking an extra-inspired ventilation. In 1971,
Rackow finally described the situation during induction of anaesthesia with the
different ventilators (81), but the different respiratory patterns he described
received little further attention and gradually faded into history.
In 1974, I came across Eger’s diagram in Scurr and Feldman’s Scientific
Foundations of Anaesthesia (22). At the time, I was a junior registrar in the
Department of Anaesthesia in the Royal Perth Hospital. The diagram appealed
to me – so simple, so elegant! It was only a year or so later that I became
suspicious of the diagram and asked Ian Ritchie if he could see any flaw in the
reasoning. He came back to me after 2 weeks and pointed out the failure to
equilibrate the extra-inspired ventilation. It took many more years to “join the
dots” and I finally presented a talk entitled “The vacuum theory of anaesthesia”
at the 1996 World Congress of Anaesthesiology in Sydney. In the talk, I focussed
on the message conveyed by Eger’s diagram, the message that there is always
an extra-inspired ventilation during nitrous oxide anaesthesia and asked the
question “How can this be when most of us use a constant volume ventilator
during anaesthesia?”. Edmond Eger and John Severinghaus 11 were both present.
By this time, I had approached Bill Mapleson for help to bring the
inaccuracy in Eger’s diagram to the attention of both teachers and authors. The
next year our collaboration finally bore fruit when our article entitled
“Concentration and second gas effects – can the accepted explanation be
improved?” was published in the British Journal of Anaesthesia 12 (48).
Following this, Eger’s diagram virtually disappeared from those anaesthetic
textbooks published outside the United States. Within the United States,
however, the diagram continued to be produced regularly by Eger until his death.
One of the few concessions he made to our criticism is contained in his chapter:
Inhaled Anesthetics: Uptake and Distribution in the 7th edition of Miller’s
Anesthesia (20). He wrote:
“This explanation has been criticized as being overly simplistic and

ignoring the realities of some aspects of ventilation. For example, if
ventilation is controlled with a volume-limited respirator, an augmentation
in inspired ventilation is limited to the period of the expiratory pause.
Spontaneous ventilation minimizes this limitation. In any event, the reader
needs to be aware that although Figure 2.4 describes the basic factors
governing the concentration and second gas effects, the actual situation is
more complex.”
Well-known Anaesthesiology research worker & inventor of the Severinghaus PCO2 electrode (83).
11
This may be viewed at:https://bjanaesthesia.org/article/S0007-0912(17)39990-7/fulltext

12
An earlier version of this paper had previously been rejected by ANESTHESIOLOGY.
Of course, the exact opposite is true. The situation is much simpler –
uptake of significant volumes of gas concentrates all remaining gases in a
smaller volume; this increases their partial pressures and accelerates their uptake.
The only complexity has been introduced by Eger himself in persisting with a
diagram that is wrong. So confused has the picture become that Calvey and
Williams, when alluding to the concentration effect in their chapter on
Inhalational Anaesthetic Agents in Principles and Practice of Pharmacology for
Anaesthetists seem too scared to actually offer an explanation. Instead they state:
“The cause of this phenomenon is obscure” (7).
American authors other than Eger continue to reproduce the diagram
uncritically. Sometimes the diagram is disguised but the explanation remains the
same as in the following example by Forman and Benkwitz (29):
Figure 2.9. The Forman & Benkwitz version of Eger’s diagram.

Reproduced with permission from the copyright owners McGraw-Hill Health Inc.
Other times the diagram remains unchanged, but the explanation is altered
as in this extract from the chapter by Ebert and Naze (17) in Clinical Anesthesia:
“In this hypothetical example, the second gas is set at 2% of a potent
anesthetic and the model is set for 50% uptake of the first gas, nitrous oxide
in the first inspired breath. The second gas is concentrated because of the
uptake of nitrous oxide (middle panel). On replenishing the inspired
second gas in the next breath, the second gas has been concentrated to be
2.7% because of the uptake of nitrous oxide in the previous breath.”
In the version of the Eger-Stoelting diagram accompanying this
explanation, the second gas concentration starts at 2% in the first panel, rises to
3.1% in the second and finishes up at 2.7% in the third. Note however, that the
dilution in the third panel is now ascribed to the next breath implying that one
respiratory cycle consists of inspiration-expiration-inspiration. Does this mean
that the next respiratory cycle consists of expiration-inspiration-expiration?
Surely this is a classic case of trying to make the facts fit the theory! One can
only feel pity for the poor anaesthetic trainee trying to make sense of such
gobbledygook.
It is often said that:
One picture is worth a thousand words.
This assumes the picture is right.
The Eger-Stoelting diagram illustrates exactly what can happen when the
picture is wrong!
A CHALLENGE TO THE VALIDITY OF THE SECOND GAS EFFECT
In 1999, Sun et al. published an article in Anesthesia & Analgesia claiming
that the second gas effect is not a valid concept (85). They studied the effects of
N2O on the ratio of end-tidal concentration to inspired concentration of 0.2%
enflurane and on its blood concentration in 14 young healthy males (7 subjects
in each group). They reported that nitrous oxide did not affect the end-tidal or
blood concentration of enflurane under controlled constant volume ventilation
and that the second gas effect is therefore not a valid concept.
Bill Mapleson contacted me soon after and invited me to join him in
writing a Letter to the Editor rejecting Sun’s claims (59).
Figure 2.10. Letter to Anesthesia & Analgesia by Mapleson & Korman.

Reproduced with permission.
In response, Lee and Sun 13 made the following points (50):
1. Disproportionate removal of one gas from a gas mixture changes the
composition of the remaining gases. Mathematically, this produces
concentration and second gas effects. These are difficult to verify
experimentally and are therefore unimportant.
2. They did not believe that they had missed marginal differences or made a
Type II error.
The message agreed on by both sets of authors seemed to be that the second
gas effect is not important clinically. 14 However there remained a clear point of
difference in that Lee and Sun maintained that the second gas effect is difficult
to verify.
The same year, Taheri and Eger demonstrated a significant second gas
effect for desflurane in 65% nitrous oxide and argued against the conclusion
presented by Sun et al. (86). A study by Nishikawa et al. published in 2000
provided evidence in 15 patients confirming the presence of a second gas effect
for oxygen in the presence of 66.7% nitrous oxide with the average PaO2
increasing from 172 mmHg to 201 mmHg (65). It was in this setting that the
work of Peyton et al. appeared and challenged existing concepts.
13
The full response may be seen at:
https://journals.lww.com/anesthesia-
analgesia/Fulltext/1999/11000/The_Second_Gas_Effect_Is_a_Valid_Concept.51.aspx
14
Although this conclusion needs to be revised given the events that followed.
VENTILATION-PERFUSION MISMATCH APPEARS AS A FACTOR
Dr Philip Peyton was completing a Cardiac Fellowship at the Alfred
Hospital in 1995 when he became interested in a project led by Dr Gavin
Robinson, one of the Staff Anaesthetists, who was investigating a new non-
invasive technique of continuous cardiac output measurement: the inert gas
throughflow technique (80). This was being compared with the standard
thermodilution method. Robinson had identified that A-a gradients for nitrous
oxide might seriously affect the accuracy of the method. It was decided to
investigate the kinetics of nitrous oxide uptake using a multi-compartment
computer model of the lung which incorporated ventilation-perfusion mismatch.
This work ultimately became the subject of Peyton’s MD thesis (66). Data from
the model predicted unexpected findings in relation to the second gas effect
[Peyton, personal communication]. These then became the subject of a separate
study which produced 3 important papers 15 in the prestigious Journal of Applied
Physiology (68-70).
In the first of the papers, a two-compartment model was used to predict
that nitrous oxide uptake increases paradoxically in the presence of V Q
mismatch. The second paper extended the analysis to a multi-compartment
model with a log normal distribution of V Q ratios and investigated the effect
on gas uptake of increasing the standard deviation of the distribution of perfusion
to mimic the values seen during anaesthesia. The results for the two-
compartment model were confirmed in the multi-compartment model. In the
third of the papers, it was shown that concentration and second gas effects may
15
The Abstracts may be viewed at: https://journals.physiology.org/toc/jappl/91/1
persist even as nitrous oxide uptake falls to the low levels reached after the initial
rapid rise of FA FI has been completed and attributed the persistence to the
increase in V Q mismatch.
Peyton directed his efforts at investigating the likely effects on the arterial
oxygen partial pressure - clearly of importance clinically. Thus, the models were
aimed at mimicking physiological conditions, so oxygen, carbon dioxide,
nitrogen, nitrous oxide and a volatile anaesthetic agent were included in the
analysis. Constant inflow and constant outflow models were studied. Absorption
atelectasis and hypoxic pulmonary vasoconstriction were incorporated when
constant inflow was being simulated.
The predictions regarding oxygen were confirmed clinically by a study in
which 30% oxygen was administered either with 70% nitrogen or 70% nitrous
oxide and the change in PaO2 measured using patients as their own controls (71).
PaO2 remained significantly higher over 45 minutes after the introduction of
nitrous oxide, consistent with the presence of persisting concentration and
second gas effects. Similar findings were obtained by Hendrickx et al. when
sevoflurane was the second gas under investigation (34).
THE SEVERINGHAUS OBJECTION
In a Letter to the Editor of the British Journal of Anaesthesia (82),
commenting on the article by Hendrickx et al. (34) in which the claim is made
that large volume N2O uptake alone cannot explain the second gas effect on
sevoflurane, John Severinghaus made the following points:
1. The findings of Hendrickx et al. conflict with theoretical analysis of the
second gas effect.
2. This casts suspicion on the accuracy of their analytical methods which
involved the use of a Datex analyser.
3. The analyser itself may have been faulty or incorrectly calibrated.
4. Alternatively, infrared analysis, the technique involved in the measurement
of sevoflurane concentrations in gas mixtures may have been interfered with
by other gases present: CO2, O2, N2O, H2O.
5. The results should have been checked by an alternative method of gas
analysis: mass spectrometry or gas chromatography
He concluded with the following paragraph 16:
“The authors speculate that pulmonary ventilation to perfusion
inhomogeneity might explain this effect, and provide some background evidence
in support. The idea is interesting, but I cannot rationalize it being so independent
of the uptake of N2O.”
Clearly, the idea of ventilation-perfusion mismatch as a cause of increased
uptake of sevoflurane did not sit well with Severinghaus, who preferred to look
for other ways of explaining the results. Moreover, this is not surprising because
it is not at all obvious how such mismatch could actually facilitate the increased
uptake. This then, was the status quo when I first became aware of Philip
Peyton’s work.
16
The full correspondence may be viewed at:
https://bjanaesthesia.org/article/S0007-0912(17)35134-6/fulltext
SUMMARY AND IMPLICATIONS
1. The Eger-Stoelting diagram is often used as a teaching tool. The steps in the
diagram do not match those taken in the mathematical model used by Eger to
predict the existence of the concentration effect.
The concentration and second gas effects are best explained as follows:
Uptake of significant volumes of gas concentrates each remaining gas in a smaller volume; this
increases its partial pressure and accelerates its uptake.
If the author feels so inclined, a second sentence may be added:

In the case of CO2, its elimination from blood is slowed.
No diagram is necessary but if one is to be included the following is sufficient.
Figure 2.11. Correct version of the Eger-Stoelting diagram.

Reproduced with permission from Korman (44)
2. Whenever large volumes of gas disappear into blood during gas uptake a
significant difference may arise between the inspired and expired tidal volume
of each breath. When modelling anaesthetic uptake with a fixed FRC, two
extreme patterns of respiration are recognisable: constant inflow and constant
outflow. Each of these exhibits the concentrating effect shown in Figure 2.8.
Any combination of these two patterns will therefore exhibit a concentrating
effect. 17
3. There is theoretical and experimental evidence suggesting that ventilation-
perfusion mismatch can amplify the concentrating effect of gas volume uptake
produced by the absorption of nitrous oxide. This may lead to the persistence
of significant concentration and second gas effects even when the net gas
volume being transferred to blood decreases to the levels seen as FA FI for
nitrous oxide reaches plateau values.
Since I was intimately involved in establishing the truth of 1 and 2 above,
the way forward in the research for this thesis now became clear. Its focus
became proposition 3. In this regard, I set out to do the following:
A. Independently verify the accuracy of Peyton’s predictions
B. Explore further the implications of his findings
C. Determine exactly how V Q mismatch acts to enhance gas uptake.
17
Furthermore, it is not possible to have constant outflow in the absence a concentrating effect in
some form or other.
2.15.1 INDEPENDENT VERIFICATION
The need to independently confirm the existence of this V Q effect is
important. From personal experience with the various methods of analysis – gas
chromatography, infra-red spectroscopy and mass spectrometry – I know how
finicky and time consuming they can be. They require a degree of skill which I
never quite acquired. Hence, I am somewhat reluctant to place too much reliance
on experimental results based on the use of these techniques. On the other hand,
I was involved in marketing MEDICHECK ANAESTHESIA, the first computer
billing program specifically written for Australian anaesthetists by my brother-
in-law Dr John Kessell (inventor of the Kessell laryngoscope blade for difficult
intubations). As a result, I know only too well how easy it is to introduce a bug
in computer programs and how difficult it can be to locate and eradicate it. Could
it be that Philip Peyton had inadvertently made a mistake which had gone
undetected? While Phil would undoubtedly have given me access to his program,
I felt that I should tackle the problem afresh in order to avoid overlooking any
mistake and then incorporating it into my own model.
In his model, Phil included oxygen and carbon dioxide as well as nitrogen,
nitrous oxide and a volatile second gas. This is understandable given that he was
studying the second gas effect on oxygen, but it added factors that may lead to
the inadvertent introduction of errors, complicate the picture unnecessarily, and
make it difficult to draw general conclusions. From previous work with Jim
Bassingthwaighte and Ranjan Dash, I was only too aware of the complex
relationship between oxygen, carbon dioxide and haemoglobin (14), and
preferred to omit these gases if possible. The distinction between constant inflow
and constant outflow seemed less important once the universal applicability of
the concentrating effect became obvious, and absorption atelectasis and hypoxic
vasoconstriction simply add another layer of complexity.
Phil eschewed the assumption of an accompanying insoluble vehicle gas
in the inspired mixture. This supposition has served numerous previous authors
well, including Kety (42), Mapleson (57), Eger (18, 21), West (89-91), Evans et
al. (26), Colburn et al. (10), and Kelman (40). I could see no reason to avoid it.
The San Diego team under the leadership of John West had produced many
mathematical treatments of gas exchange in the 1970s, based on this
assumption 18.
John Evans, the professional mathematician with a medical background
who joined West’s team, had done much to formalize the mathematics of gas
exchange (25). Even today, his elegant treatments are a pleasure to behold!
Unfortunately, their beauty was probably not fully appreciated at the time by
Respiratory Physiologists with scant mathematical training, and their
significance has been lost over time. Looking over some of these papers, I
realized I could harness the power of Evans’s mathematics to study the second
gas effect, his derivation of the log normal distribution of V Q ratios could be
utilized as a framework for further investigations, and the conversion of sums to
Lebesgue-Stieltjes integrals could reduce the need for mind-numbing repetitive
calculations. I use his equations for the log normal distribution repeatedly but
freely admit that I cannot follow the steps taken in their derivation. Neither can
18
The team achieved the equivalent effect by assuming that there is no net transfer of gases other than
the one under immediate consideration. “If we imagine a lung in which the test gas is accompanied by
only one other gas, the error in this assumption can be made arbitrarily small by choosing the
solubility of the vehicle gas to be sufficiently low” (10).
any of the mathematicians I have consulted on this question. This missing
information is something of an indictment on the editorial process of the Journal
of Applied Physiology. Evans passed away in 2016, before I was able to discuss
the matter with him and the derivation is probably now lost forever. The steps
that can be determined are shown in Appendix A, at the end of this thesis. They
were kindly provided by Dr. Snezhana Abarzhi, Professor of Applied
Mathematics at the University of Western Australia.
2.15.2 FURTHER INVESTIGATION

Peyton’s prediction that gas uptake can be facilitated by an increase in
V Q mismatch is revolutionary! It is surprising that no Respiratory Physiologist
has seen fit to investigate the matter. With Leon Farhi passing in 2003 and Albert
Olszowka retired, the research school at the University of Buffalo is no longer
focused on mathematical aspects of respiratory physiology. John West is now in
his nineties and his co-worker Peter Wagner involved in developing a vineyard.
With the main players out of the picture, it has been largely left up to the
anaesthetic fraternity to examine the implications of Peyton’s unexpected
findings. I have tried to do my part, mainly by revisiting publications by West’s
San Diego group from the period in the 1970s when the Multiple Inert Gas
Elimination Technique was being developed.
2.15.3 SATISFACTORY EXPLANATION
This is the final task I set myself. What use is it to predict numerous
possibilities if one cannot understand the underlying mechanism, worse still, if
one doesn’t actually believe the phenomenon itself is possible? Having always
possessed a certain degree of scepticism, I found it hard to believe that
ventilation-perfusion mismatch could do anything other than interfere with
efficient gas exchange. If I were to succeed in confirming Peyton’s findings and
extending them, then surely the next step would have to be the search for a
satisfactory explanation. I leave it to the reader to decide whether I have been
successful.
: The Log-Gaussian Distribution
INTRODUCTION
In this chapter we deal with equations derived by Colburn, Evans and West
(10) which give rise to the density functions for ventilation and blood flow as a
function of log V Q( ) and are central to the mathematical treatment of the
concentration and second gas effects developed in this thesis. This paper seems
to be the only one published by the San Diego research team involving Colburn.
Most likely the key figure in deriving the expressions was John Evans, the
professional mathematician recruited by John West. Evans co-authored most of
the mathematical papers produced by the team and was clearly instrumental in
providing a sound theoretical basis for the Multiple Inert Gas Elimination
Technique (see  6.1).
The following equations were derived for the expired alveolar ventilation
VA , and the pulmonary end-capillary blood flow Q (see Appendix A):
V
V = 0 e − x /2
2
(3.1)
2π
Q − x −σ 2 /2
Q = 0 e ( ) (3.2)
2π
Here V0 and Q 0 are used as alternative symbols for the expired alveolar
ventilation VA and the total pulmonary end-capillary blood flow Q t , both
expressed in units of l/min. These symbols are useful when moving from a
continuous to a discrete (compartmental) description of gas exchange in the lung
Chapter 3: The Log-Gaussian Distribution 39

where they can be used to store the sum of the compartmental values. I use both
sets of symbols interchangeably in this thesis.
CHOICE OF THE LOG NORMAL DISTRIBUTION
West and Wagner (92) give several reasons why a log normal distribution
is used most often, as in Eqs. 3.1 and 3.2. Such distributions are common in
normal and abnormal lungs. It is a simple distribution whose spread is
characterised by a single parameter - its standard deviation (SD). Unlike a linear
normal distribution, serious impairment of gas exchange is possible without
developing a negative ventilation or blood flow.
CORRECT ORIENTATION OF GRAPHS
If we were to plot V and Q in Eqs. 3.1 and 3.2 against x, the graph for Q
would lie to the right of the graph for V which does not conform to the usual
way these graphs appear in the literature. However, Colburn et al. point out that
the degree of mismatch is unaffected if we substitute -x for x in the equations. If
we do this, the correct order of the graphs is restored.
THE MEANING OF σ
In Eq. 3.2, σ specifies the degree of mismatch between ventilation and
blood flow. It is derived from the standard deviation of the log ventilation per
unit lung volume, σV and the standard deviation of the log blood flow per unit
lung volume, σ Q as follows:
σ σV − σ Q
= (3.3)
and has the property of locating the position of the means for ventilation and
blood flow, rather than being a standard deviation (see  3.8).
40 Chapter 3: The Log-Gaussian Distribution

MEASUREMENTS OF σ IN ADULT SUBJECTS
In awake healthy adults, σ varies from 0.25-0.5; worsening of V Q
matching is seen in disease states and is also known to occur soon after induction
of anaesthesia, with typical values for σ between 0.75 and 1.75 in healthy adults
(16, 52-54, 78). The changes during anaesthesia are especially relevant given
that the persistent increases of the second gas effect in blood during nitrous oxide
anaesthesia have been attributed by Peyton et al. to an increase in V Q
mismatch (68-70).
THE COMPARTMENTAL APPROACH
Although Eqs. 3.1 and 3.2 allow ventilation and perfusion to each be
described as continuous functions of the ventilation-perfusion ratio (specifically
as continuous log normal distributions), this route does not appear to have been
followed by either the group of respiratory physiologists working in San Diego
under West, or the group in Buffalo under Hermann Rahn. There are several
reasons why the compartmental approach seems to have been preferred when
investigating the effects of V Q mismatch. First and foremost is the fact that
their primary interest lay in describing the exchange of oxygen, carbon dioxide
and to a lesser extent, nitrogen. The complicated algorithms defining O2-CO2
exchange are most easily dealt with by treating gas exchange in the lung as
taking place in a collection of parallel compartments and allowing each
compartment to contribute to the output in proportion to its share of the
ventilation or blood flow. West and Wagner give a second reason: “the division
into compartments is convenient and seems justified by the essentially
compartmental nature of the real lung” (92). In other words, the use of a
compartmental approach should help student and research worker visualize

events more readily. This is illustrated in Figure 3.1 in which the method of
constructing the compartments with a log normal distribution of ventilation-
perfusion ratios is shown.
Figure 3.1. Distributions of ventilation-perfusion ratios.

The distributions of ventilation per unit volume and blood flow per unit volume in (a) and
(b) are combined to give a distribution of ventilation-perfusion ratios in (c). Based on
original material by West & Wagner (92). Permission was given by JBW.
In Figure 3.1(a), the volume of gas is shown as a function of ventilation
per unit volume. This is assumed to follow a log normal distribution with mean
value μV and SD σV and has been drawn using an expression of the form:
1  ( x − μ )2 
V
=yV exp −  (3.4)
2πσV2  2σV
2

In Figure 3.1(b), the volume of gas is shown as a function of blood flow
per unit volume. This is assumed to follow a log normal distribution with mean

value μQ and standard deviation σ Q and has been drawn using an expression of
the form 19:
 x−μ
( ) 
2
1  Q 
yQ = exp −  (3.5)
2πσQ2  2σQ2 
 
A suitable partition of the x-axis allows a series of compartments to be set
up which effectively covers the whole range of values of the ventilation per unit
volume and blood flow per unit volume. By matching compartments going from
left to right across Figure 3.1(a) and Figure 3.1(b), it is possible to produce the
distributions of ventilation and blood flow as functions of log (V Q ) shown in
Figure 3.1(c). Essentially, this involves plotting the average value of yV and yQ
for each compartment as a function of the simultaneous value of log yV yQ . 20 ( )
The ease of applying this technique is probably a third reason it was favoured.
V AND Q AS EXPLICIT FUNCTIONS OF log(V/Q) 21
In applying Eqs. 3.1 and 3.2, West makes no attempt to derive explicit
( )
expressions for V and Q in terms of log V Q (91). Colburn et al. also fail to
progress this possibility further (10). This will now be investigated in more
detail.
19
Interestingly, the first negative sign inside the curly brackets in Eq. 3.4 was omitted in the original
paper by West (91). The mistake was repeated in a chapter by West and Wagner in a reference text
(92). The same error features in a scene in the film Gifted, about an 8-year-old girl set a mathematical
problem involving the incorrect expression.
20
This involves an application of the trapezoidal rule (4,36).
21
The equations presented here and in the next section are modified versions of those appearing in
reference (45). Those in  3.8 have been corrected. In the reference, v and q are each shown as
functions of time whereas they are in fact dimensionless.

Dividing Eq. 3.1 by Eq. 3.2 and taking logarithms of each side we obtain:
( Q
log V= ) ( )
log V0 Q 0 + σ 2 2 − σ x (3.6)
Solving for x:
σ
x=
2
( )
+ (1 σ ) log V0 Q 0 − (1 σ ) log V Q ( ) (3.7)
We now introduce two new variables v and q which are the fractional
ventilation V V0 and fractional blood flow Q Q 0 respectively. Letting
z = log ( v q ) Eq. 3.7 can then be written in the more compact form:
σ z
x= − (3.8)
2 σ
Substituting for x in Eqs. 3.1 and 3.2 we obtain (see also footnote 23):
2
 σ2
V −  z − 
2 
2σ 2
V = 0 e  (3.9)
2π
2
 σ2
Q −  z + 
2 
2σ 2
Q = 0 e  (3.10)
2π
( ) ( )
since z log V Q − log V0 Q 0 , a plot of V or Q as a function of z
Note that=
is identical to a plot of these variables as a function of log (V Q ) but is shifted
(
log V0 Q 0 ) units to the left if V0 > Q 0 and to the right if V0 < Q 0 . Plotting
against z has the advantage of always locating the graphs for σ = 0 at zero. When
σ > 0, the mean for V is situated exactly σ 2 2 units to the right of 0 and that
for Q exactly σ 2 2 units to the left of 0. Typical graphs of V or Q against z
can now be drawn and are shown in Figure 3.2 for values of σ = 0, 0.5, 1 and 2.

With σ = 0, the graphs are vertical lines positioned at z = 0, the blue line
for ventilation represents a flow of 4 l/min, the red line for blood flow represents
a flow of 5 l/min. As σ is increased, the graphs assume a bell-shaped curve, with
that for ventilation moving to the right and that for blood flow moving to the left.
Three changes are taking place simultaneously. First, the graphs for
ventilation and blood flow are moving away from each other in proportion to the
square of σ . This is indicated in Figure 3.2 by the vertical dash-dotted lines
Figure 3.2. Ventilation and blood flow as functions of a new variable.

Graphs of ventilation and blood flow as functions of z = log (V! Q! ) - log (V!0 Q! 0 ) for
increasing degrees of V! Q! mismatch (clockwise from top left σ = 0, 0.5, 1 and 2). Scales
on the z and y axes have been kept constant to allow the effect of changes in σ to be seen
easily.
located at the maximum value of ventilation (blue) and blood flow (red), in those
cases where σ is greater than 0. Secondly, the spread or dispersion of the graphs
is increasing in proportion to σ . Thirdly, since the area under each ventilation

curve must equal V0 and that under each blood flow curve must equal Q 0 , 22 the
height of each graph must decrease proportionally as σ is increased. 23 These
changes make it difficult to detect and interpret alterations in other variables such
as gas uptake and to relate these in a meaningful fashion to the underlying
changes in V Q as the degree of mismatch is increased.
v AND q AS EXPLICIT FUNCTIONS OF (1/σ)log(v/q)
A more useful picture can be obtained by plotting against
x = (1 σ ) log ( v q ) on the x-axis. This unit is suggested by Eq. 3.7. When this
substitution is made in Eqs. 3.9 and 3.10 we obtain:
2
V0 −  x − 2 
 σ 
2

V= e (3.11)
2π
2
Q 0 −  x + 2 
 σ
2

Q= e (3.12)
2π
If in addition, v and q are plotted on the y-axis instead of V and Q , this
results in the graphs for ventilation and perfusion being the same shape and size
for every value of σ . All that changes as σ is increased is the location of these
graphs relative to the y-axis and each other. The maximum for v is located
exactly σ 2 units to the right of the y-axis while that for q is located exactly σ 2
units to the left of the y-axis.
22
In this thesis, we generally set these at 4 l/min and 5 l/min respectively.
23
The scaling occurs because of the change of variable, so when integrating Eqs. 3.1 and 3.2 using the
substitution shown in Eq 3.8, dx = − ( d z σ ) .

Figure 3.3. Distributions of fractional ventilation (v) and blood flow (q).
Distributions of fractional ventilation and fractional blood flow as a function of
x = (1 σ ) log ( v q ) for increasing degrees of V Q mismatch. Clockwise from top left σ
= 0. 0.5, 1, 2. The distributions are now symmetrical about the vertical line at x = 0. The
curve for ventilation is blue; that for blood flow is red. All curves now are identical in
shape and size and have a standard deviation of 1 unit. The means for ventilation and
blood flow are located σ 2 units to the right and left of the y-axis (blue and red dash-
dotted lines respectively). When σ = 0 , the curves overlap. The area under each curve is
equal to 1.
The maximum value of v and q is 1 2π which is approximately equal to
0.4. This is also the mean value for the distribution. 24 The results are shown in
Figure 3.3. The process of gas exchange can be visualized further by rotating the
graph for blood flow by 180° around the x-axis. This is shown in Figure 3.4 for
σ = 0, 0.5, 1 and 2.
Consider now the case when σ = 2 in Figure 3.4. To the left of the vertical
dash-dotted lines at − σ 2 , there is a paucity of ventilation and an excess of blood
24
It can be seen that σ has the property of locating the position of the mean as stated in  3.2 above,
rather than being a SD. In fact, we use the symbol μ for σ 2 in parts of this thesis.

flow. Here ventilation acts as an impediment to gas exchange. To the right of the
vertical dash-dotted lines at σ 2 , there is a paucity of blood flow and an excess
of ventilation. Here blood flow acts as an impediment to gas exchange.
Figure 3.4. Distributions of fractional ventilation (v) and blood flow (-q).
Fractional ventilation (v) and fractional blood flow (q) distributions as a function of
x = (1 σ ) log ( v q ) for increasing degrees of V Q mismatch with σ = 0, 0.5, 1, and 2
(clockwise from top left). The curve for ventilation is blue; that for blood flow is red. The
curve for blood flow has been reflected in the x-axis to show the effect of mismatch more
clearly. The area between each curve and the x-axis is equal to 1.
This interpretation is specific to a log normal distribution of ventilation and
blood flow ratios but conforms with the classical view that V Q mismatch must
inevitably impede gas exchange in the lung.

THE CASE WHEN σ = 0
When ventilation and perfusion are perfectly matched, i.e. σ = 0, the red
curve for blood flow in Figure 3.4 is the mirror image of the blue curve for
ventilation. This is sufficiently different from the situation in Figure 3.2A in
which ventilation and blood flow are located on a single vertical line, as to
require further explanation. When σ = 0, V = V0 and Q = Q 0 , so
(1 σ ) log ( v q ) = 0 0 and is therefore indeterminate. The situation must be
treated by finding lim (1 σ ) log ( v q ) , which from Figure 3.4 clearly results in
σ →0
the red and blue vertical lines approaching each other more and more closely
until they coincide at x = 0.
An alternative approach is to consider the case when σ is very small but
not exactly equal to zero, say σ = 0.001. Clinically, we would not be able to
distinguish the situation from the case when σ actually equals 0 but
mathematically we have a log normal distribution with a very narrow standard
deviation of 0.001. When we divide by 0.001 to obtain the associated x value,
the values of log ( v q ) are increased a thousand fold to produce curves of the
same size and shape as those in Figure 3.3 and Figure 3.4. At each x-value, v =
log ( v q ) =
q, so z = ( ) (
log V Q − log V0 Q 0 = )
0 giving the required condition
that V Q = V0 Q 0 .
INTERPRETATION OF GAS EXCHANGE USING FIGURE 3.4
When σ is greater than 0, any point to the right of x = 0 is associated with
a ventilation-perfusion ratio in excess of VA Q t while any point the left of x = 0
is associated with a ventilation-perfusion ratio less than VA Q t . Thus we

preserve the notion of high V Q gas-exchanging units to the right of VA Q t and
low V Q gas-exchanging units to the left. If we wish to show actual gas uptakes
on the y-axis, all that is necessary is to multiply the fractional ventilation v, by
VA and the fractional blood flow q, by Q t . In this thesis, we will often be using
diagrams with (1 σ ) log ( v q ) as the variable on the x-axis as a platform on
which to study the uptake of anaesthetic gases. Any questionable results can be
( )
investigated further by reverting to log ( v q ) or log V Q as the independent
variable.
INTERCHANGEABILITY OF X-VARIABLES
So far, we have been interested in the distributions of ventilation and blood
flow V and Q , and the derived variables v and q. Traditionally, these have been
plotted as functions of log (V Q ) or log10 (V Q ) . In later parts of this thesis we
are also interested in the distribution of gas uptake i.e. where gas uptake occurs
in the V Q spectrum. The question arises as to whether a trend evident when
we plot these distributions as a function of our new variable (1 σ ) log ( v q ) ,
mirrors a similar trend in the distribution when plotted against log (V Q ) . In
other words, are we likely to miss some trend by using the new variable on the
x-axis? This question is best answered as follows.
Assume we have a variable whose distribution D is a function f of y where
y = (1 σ ) log ( v q ) i.e. D( y ) = f ( y ) . Assume further that the distribution may
also be expressed as a second function g(x) where x = log (V Q ) i.e.
D( x) = g ( x) . A trend in D(x) is given by dD dx , while a trend in D(y) is given

by dD dy . Since dD dx = ( dD dy ) . ( dy dx ) , and ( dy dx ) = 1 σ , where σ ≥ 0
, it follows that a trend in D(y) does indeed reflect a similar trend in D(x).
Similarly, a trend in D(x) is mirrored by a corresponding trend in D(y).
LEBESGUE-STIELTJES INTEGRATION
Figure 3.5. Partitioning the x-axis using a compartmental approach.
As we have mentioned in  3.6, the standard method adopted by West and
his colleagues when equilibrating alveolar gas with pulmonary capillary blood
is to set up a partition on the x-axis to mimic the assumed compartmental nature
of the lung. The principle is illustrated in Figure 3.5. Equilibration between gas
and blood is then performed on a compartment-by-compartment basis and the
output for each phase calculated by adding together the appropriately flow-
weighted contributions from each compartment.
Looking at Figure 3.5, it is difficult to avoid a comparison with the
standard technique of Riemann Integration taught in High School calculus
courses, in which the area under a curve is taken as the integral of the function
involved in generating the curve and is approximated more and more closely by

finding the limit of the sum of the subtended areas under the curve as the partition
of the x-axis is made finer and finer (4). It should therefore come as no surprise
to learn that the appropriate integral, known as the Lebesgue-Stieltjes integral,
has been described in the case of a log normal distribution of V Q ratios by
Colburn et al. (10) although there is no evidence that it was put to use by them
to solve problems in Respiratory Physiology.
The theoretical basis of the Lebesgue-Stieltjes integral is explained in
many mathematical texts e.g. McShane (60). While the theory is somewhat
complicated, the application of the integration process in the case of the
compartmental models of gas exchange we deal with in this thesis is relatively
straightforward. If some variable say Gi in the ith compartment is a function of
Vi and Q i i.e.:
(
Gi = f Vi , Q i ) (3.13)
and the Gi are summed to obtain the output variable G i.e.:
n
G = ∑ f (Vi , Q i ) (3.14)
i =1
then in the limit as n → ∞ the summation may be replaced by an integral in
which Vi and Q i are replaced by expressions given in Eqs. 3.1 and 3.2 or variants
of these expressions shown above, for example:
∞
 V −  x − σ 
2 2

Q 0 −  x + 2 
n  σ
( ) ∫
2 2
=G lim
= ∑  
f Vi , Qi f  0 e  2

, e  dx

(3.15)
−∞  2π 2π
n →∞
i =1

Both the summation (West’s method) and Lebesgue-Stieltjes integration
should yield identical results if n is sufficiently large in the former. In practice,

we use whichever technique is most convenient in dealing with any particular
problem.
SUMMARY
In this chapter we have developed a framework for studying the SGE in
the presence of V Q mismatch. This involves assuming a normal distribution
for ventilation and blood flow as functions of the logarithm of a suitable variable.
The independent variable chosen by Respiratory Physiologists in the past is
( )
log V Q , the logarithm of the V Q ratio. We have proposed an alternative
independent variable (1 σ ) log ( v q ) , where v and q are the fractional ventilation
and fractional blood flow respectively. The use of this variable provides a stable
platform for the graphical interpretation of gas exchange in later chapters of this
thesis. The suitability of this variable lies in the fact that the SD of the curves for
ventilation and blood flow is always 1 regardless of the value chosen for σ .
Moreover, the degree of V Q mismatch is controlled by a single parameter σ ,
whose values are known in various clinical situations. The effect of V Q
mismatch on various aspects of gas exchange can then be studied graphically,
by varying σ in a stepwise fashion on a stable platform.
In the 1970s, West and Wagner described the standard method of treating
such normal distributions by subdividing the continuous distributions into a
number of parallel gas-exchanging compartments but noted that “it is possible
to dispense entirely with the compartments and treat the distribution as a density
function. Such a mathematic maneuver enables one to make general statements
about the behaviour of the distribution” (92). This possibility was not
investigated further by them. In this chapter, we have laid the groundwork for

using the technique of Lebesgue-Stieltjes integration as a preferred option
whenever possible in this thesis.

: Steady-state Equations
In this chapter we derive the equations used to study the exchange of the
first and second gases in any gas-exchanging unit of the lung, assuming that
steady-state conditions apply. We develop the simplest possible model for
studying the second gas effect. This involves excluding any factor which does
not contribute significantly to the phenomenon and, from our perspective,
therefore does not play an essential role in its production. Thus, we exclude the
physiological gases oxygen and carbon dioxide from consideration. Clearly, they
may themselves be the subject of a second gas effect and can also make a small
contribution to the net gas volume change during nitrous oxide anaesthesia.
However, their interaction with haemoglobin is so complex that their inclusion
will necessarily complicate any mathematical treatment, potentially obscuring
important results that would otherwise be obvious. This simplification has been
used by many previous workers in the field (18, 21, 42, 57, 58). Absorption
atelectasis (12), which was included in the modelling of Peyton et al. (68-70),
has also been excluded as it is not an essential ingredient in the production of the
second gas effect. The equations being sought are those describing the fractional
concentration, partial pressure and uptake of the gases involved. These are
elaborated further in subsequent chapters to obtain specific relevant results. The
choice of a steady-state model is an important step and will now be discussed
further.
THE NON STEADY-STATE MODEL
The uptake of anaesthetic gases necessarily involves a non steady-state.
Thus, during induction of anaesthesia, existing alveolar gas is replaced by the
Chapter 4: Steady-state Equations 55

anaesthetic gas whose solubility in blood is λ and whose inspired concentration
is FI . 25 If the inspired and expired alveolar ventilations in a lung with no V Q
mismatch are respectively VIA and VEA , 26 then a mass balance equation of the
following form will apply:
VIA FI − VEA FA − d (VL FA ) dt = λQ ( Fa − F v ) (4.1)
Here FA is the alveolar concentration at time t, VL is the alveolar volume,
and
= ( , F v Pv
Fa Pa PB − PH 2O= ) (P
B )
− PH 2O , 27 where Pa , P v , PB , P H 2O are
the arterial partial pressure and mixed venous partial pressure of the gas,
barometric pressure and saturated vapour pressure of water at 37 degrees
centigrade respectively. Equations of this type were solved first with VIA = VEA
by Kety (42) and Mapleson (57), and then with differing inspired and expired
ventilations by Mapleson (58) and others.
Attention is drawn to the term d (VL FA ) dt which describes the change in
volume of anaesthetic in alveolar gas as a function of time. When expanded
using the chain rule, this gives rise to the terms VL FA + FAVL . Although the
functional residual capacity (FRC) is now known to decrease during anaesthesia,
this was not the case when the equations were first solved so the term FAVL was
assumed to be equal to zero. Even now, there is no readily available
mathematical description of this process, so the term is generally ignored. This
25 Concentrations are defined by convention as fractional concentrations in dry gas and may be converted to
partial pressures through multiplication by (P − P ) .
B H 2O
26 By convention VA = VEA , the expired alveolar ventilation.

27 Fa and F v may be thought of as the concentrations in the dry fraction of a gas sample equilibrated with
arterial or mixed venous blood respectively.
56 Chapter 4: Steady-state Equations

leaves the term VL FA which reduces Eq. 4.1 to a first order differential equation
which is usually solved by computer. Non-linearity is introduced whenever VEA
is allowed to vary with constant inflow and also results from the varying inputs
to P v from various tissues and organs.
Each gas present will have its own mass balance equation. Those for
oxygen and carbon dioxide are more complicated because there is no single
value for λ , as a result of the complex interaction with haemoglobin. There is
also a mass balance equation for the net gas volume change. All of these
equations must be solved simultaneously if a full description is to be obtained.
Clearly, we are dealing with a very complicated system to which we now
wish to add an investigation of the effect of V Q mismatch. To the best of my
knowledge, no model of anaesthetic uptake has been solved to date which
includes an accurate description all of these factors.
THE STEADY-STATE MODEL
A practical alternative is to eliminate the d (VL FA ) dt term completely
from consideration. This approach is not new and has been adopted by previous
workers in the field (10, 27, 48, 68-70). Several benefits follow immediately:
1. Calculations using Eq. 4.1 are greatly simplified.
2. The composition of gas in the FRC remains constant from the end of one
breath to the end of the next so that FRC gas can be completely ignored and
the inspired alveolar tidal volume equilibrated directly with the pulmonary
blood flow in much the same fashion as was done by Riley and Cournand
(79). This also applies to individual gas-exchanging units in the lung.

The idea of using steady-state conditions is not entirely inappropriate.
These are approached as the alveolar concentration plateaus during induction
and may also be regarded as applying over very small periods of time during the
initial phase of rapid rise. Both phases are shown in Figure 4.1.
Figure 4.1. Washin curve for several anaesthetic gases.

After a rapid initial rise in the ratio of alveolar concentration to inspired concentration
each curve tends to plateau. (Curves derived using the model in Appendix 2).
Furthermore, Peyton has shown that a “square root of time” rule applies to
the change in uptake of all gases and volatile agents (66). This is consistent with
Severinghaus’s formula (83) and led Peyton to conclude that “after 5 minutes,
the rate of change of uptake of any gas will be only 10% of its value at that time
and that this declines to 5% after 10 minutes and 1% after 50 minutes”. This low
rate of change provides further support for the use of the steady-state approach.

By choosing appropriate values for the inspired and mixed venous partial
pressures of the first gas, (generally nitrous oxide in most simulations in this
thesis), it is possible to mimic the net gas volume changes occurring at any
specified stage of anaesthetic induction. Since the concentration and second gas
effects are generated by the net change in gas volume and do not depend on
whether we are dealing with steady-state gas exchange or not, and given the
other advantages mentioned above, the choice of a steady-state model is
especially attractive and will be used almost exclusively throughout this thesis.
A SUITABLE MODEL TO STUDY THE SECOND GAS EFFECT (SGE)
The minimum number of components in the gas phase necessary for a
theoretical study of the SGE are as follows:
1. A first gas (FG). Alveolar-capillary uptake of this gas generates the volume
changes responsible for the SGE. It must be soluble in blood and should
comprise a significant portion of the inspired ventilation.
2. A second gas (SG). This is usually a volatile agent used in low
concentrations. This gas responds to the volume changes by exhibiting the
SGE. We use the subscript SG/FG to indicate that the SG is used together
with the FG.
3. The remaining gas. This acts purely as a vehicle and for the purposes of the
analysis is regarded as being completely insoluble in blood 28.
As the steady-state equations governing the exchange of the first gas in a
single lung unit, we take:
28
As mentioned in a previous footnote, the San Diego team usually achieved the same outcome by
specifying no net exchange of the remaining gases (10, 27). In our 3-gas system only the first gas is
permitted to contribute significantly to gas volume contraction.

VIFI − VAF = λQ ( F − F v ) (4.2)
and for the insoluble carrier gas:
(1 − FI )VI =
(1− F ) VA (4.3)
where
VI = incoming alveolar ventilation per unit time (the “inflow”)
VA = outgoing alveolar ventilation per unit time (the “outflow”)
Q = blood flow per unit time
FI = concentration of the first gas in the dry fraction of the inflow
F v = concentration of the first gas in the dry fraction of a sample of gas
equilibrated with the incoming mixed venous blood
F = concentration of the first gas in the dry fraction of the outflow which is also
in equilibrium with the outgoing arterial blood
λ = Partition coefficient of the first gas in blood at 37 deg C
To this system we now add a second soluble gas, denoted by the subscript
“SG” to allow it to be distinguished from the first gas. (In the absence of a “FG”
or “SG” subscript, the variable refers to the first gas.) The second gas also obeys
Eq. 4.2 above:
VIFI SG − VAFSG = λQ ( FSG − F vSG ) (4.4)
Based on the conservation of mass, Eqs. 4.2 and 4.4 state that the amounts
of the first and second gases leaving the gas phase must equal the amounts of
these gases gained by the blood phase. The remaining gas is treated as if it were

completely insoluble, and merely serves as a “vehicle” for delivering the first
and second gases. Equation 4.3 expresses the conservation of the remaining gas.
The second gas is assumed to be present in such a low concentration that its
contribution to the net volume change may be ignored. Thus, it does not appear
in Eq. 4.3. For convenience, we set F vSG to zero, effectively excluding
recirculation. This maximizes the uptake of the second gas making the detection
of the SGE easier. Clinically, this is equivalent to the administration of minute
traces of second gas at intermittent intervals. From Eq. 4.4 we then have:
VI FISG
FSG = (4.5)

VA + Qλ SG
4.3.1 HOMOGENEOUS LUNG

First Gas in gas phase
This case is obtained by solving Eq. 4.3 for VI and eliminating this
variable from all equations.
 1− F  
VI =   VA (4.6)
 1 − FI 
Substituting for VI in Eq. 4.2 we may solve this equation for F .
VI FI + λQ (1 − FI ) F v
F= (4.7)
VA + λQ (1 − FI )
Second Gas in blood phase
We denote the uptake of the first gas N2O by the symbol VN2O , the uptake
of the second gas in the absence of the first gas by VSG and the uptake of the
second gas in the presence of first gas as VSG / FG . VN2O is given by the left-hand
side of Eq. 4.2; VSG is given by the left-hand side of Eq. 4.4 when VI = VA ; and

VSG / FG is given by the left hand side of Eq. 4.4 with VI determined from Eq. 4.6 .
Using the superscript “hl” for a homogeneous lung:
λVA( FI − F v )
VNhlO = (4.8)
2
λ(1 − FI ) + (VA Q )
 VA λSG FISG 

VSGhl =  (4.9)
  
 λSG + (VA Q) 
 VA λSG FISG   1 − F 

VSG/FG
hl
= (4.10)
   
 λSG + (VA Q)   1 − FI 
Note that Eq. 4.8 was derived by Colburn, Evans and West (10). Except
for the appearance of the factor (1 − F ) (1 − FI ) and the omission of the F v
term, Eq. 4.10 is similar to Eq. 4.8. Because FI SG  1 , it has been omitted from
the denominator of Eqs. 4.9 and 4.10.
4.3.2 NON-HOMOGENEOUS LUNG

Consider now a non-homogeneous lung model, denoted by the superscript
“nhl”, composed of n functional units. Let Vj and Q j be the alveolar ventilation
and blood flow, respectively, to the jth compartment for 1 ≤ j ≤ n. Set
V1 +  + Vn and Q=

V0 = 0 Q1 +  Q n . 29
SG in blood phase
By analogy with Eq. 4.10, the net uptake of the second gas in the presence
of the first gas in the non-homogeneous lung, VSGnhl/ FG , is given by:
29
As was pointed out in  3.1 we use the terms V0 & VA and Q 0 & Q t interchangeably, where VA
refers here to the total expired alveolar ventilation per minute.

n  V λ FI   1 − Fj 
VSG
nhl
/ FG = ∑ 
j SG SG
     (4.11)
j =1  λSG + V j Q j   1 − FI 
By analogy with Eq. 4.9, the net uptake of the second gas in the absence
of the first gas in the non-homogeneous lung, VSGnhl , is given by:
n  V λ FI 
VSG
nhl
= ∑  j SG SG  (4.12)
  
j =1  λSG + V j Q j 
SUMMARY
In this chapter, we have shown how it is possible to obtain expressions for
a number of output variables to be used in later parts of this thesis. These
variables include the mixed alveolar concentrations and partial pressures of the
first and second gases, the rate of uptake by pulmonary capillary blood of these
gases and the partial pressures of the gases in arterial blood. The derivation has
involved a number of assumptions, the most important of which is the
assumption that gas exchange is taking place under steady-state conditions, a
necessary assumption given the complexity of the system under study.

: Persistent Second Gas Effects
In this chapter we investigate the persistence of second gas effects as
nitrous oxide uptake falls to maintenance levels 30. By maintenance levels we
mean the levels associated with plateauing of the washin curve for nitrous oxide
shown in Figure 4.1. The value of 100 ml/min has been given by Severinghaus
(83).
The existence of this phenomenon was first predicted by Peyton et al. (68),
using a computer model and was subsequently confirmed experimentally for
oxygen and sevoflurane (34, 67, 71, 72). The explanation given was that the
persistence was due to the increased degree of V Q mismatch which occurs
routinely during anaesthesia (52-54, 78). This was questioned by Severinghaus
(82) who believed that the SGE cannot be divorced from significant nitrous
oxide uptake (see  2.14). The idea that a phenomenon traditionally believed to
be caused by the large reductions in gas volume associated with high N2O
uptake, can persist at clinically significant levels even when those volume
changes fall away, is a significant departure from traditional teaching.
Since 2006, there has been no real progress in resolving this dispute. The
current chapter attempts to provide useful answers. We set out to further explore
the mechanism of this proposed phenomenon with computer modelling of
theoretical distributions of V Q scatter seen in patients under general
The material in this chapter is used with the permission of ANESTHESIOLOGY and Wolters Kluwer
30
Health Inc, copyright owners of the published material.
Chapter 5: Persistent Second Gas Effects 65

anaesthesia, across a range of N2O uptake rates, and the predicted effect on
second gases with different solubilities in blood.
The sequence adopted in this chapter is to define the Washin Ratio (WR),
which we use to simulate the degree of nitrous oxide washin. We need this
measure to reliably investigate persistent SG effects during the plateau phase of
nitrous oxide washin. We next define our measure of the SGE, the Augmentation
Ratio (AR). Equations for the augmentation ratio are then derived using results
obtained in Chapter 4, and solved for typical steady-state situations during
anaesthesia.
THE WASHIN RATIO
The proportion of nitrous oxide washin is expressed as a washin ratio (WR)
where:
VN2O
WR = 1 − (5.1)
VI FI N2O
This ratio is to be used as an independent variable on the x-axis. Note that
when VN2O = 0, WR = 1, and nitrous oxide uptake is complete. When
VN2O = VI FI N2O , its maximum possible value, WR = 0, a situation that can only
apply at the commencement of anaesthesia (see Table 5.1).
Table 5.1. The nitrous oxide Washin Ratio
Stage of
Washin Ratio N2O Uptake FA FI
Anaesthesia
Start of washin 0 VI FI N2O 0
End of washin 1 0 1
66 Chapter 5: Persistent Second Gas Effects

Since we are interested in studying SGE persistence during the
maintenance phase of nitrous oxide anaesthesia we restrict our attention to values
of VN2O for which 0.9 ≤ WR ≤ 1.0 , i.e. when washin is 90 to 100% complete.
Applying Severinghaus’s equation for the rate of uptake of nitrous oxide (83),
this should occur within 10 to 15 minutes from the start of induction, at which
time the rate of change in the uptake of nitrous oxide ( VN2O ), has declined to less
than 5% of its current rate of uptake. The washin ratio is roughly equivalent to
the alveolar/inspired concentration ratio FA FI shown in Figure 4.1.
THE AUGMENTATION RATIO
We need to be able to quantify the second gas effect if comparisons are to
be made during different stages of anaesthesia. The SGE is expressed as an
augmentation ratio (AR), defined as the fractional increase in concentration of
second gas in the presence of N2O, over that which would have existed in the
absence of N2O, with all other variables kept constant. With no SGE, the
augmentation ratio is 1. Once the SGE becomes evident, the augmentation ratio
exceeds 1. The term “augmentation ratio” was first used by Epstein, Rackow,
Salanitre and Wolf (24). Although they only calculated augmentation ratios in
end-tidal gas, the ratio may also be calculated for blood. The ratio is defined as
follows:
Value of parameter in presence of nitrous oxide

AR= (5.2)
Value of parameter in absence of nitrous oxide
The parameter may be a partial pressure, concentration or gas uptake.
The advantage of the AR is the ease with which it can be applied in
experimental situations. It does however have certain drawbacks. One obvious

limitation is the inability to use it in relation to the first gas, in this case nitrous
oxide, and hence it cannot be used to measure the concentration effect. Other
disadvantages are discussed elsewhere in this thesis (see  7.17). In this chapter
we obtain equations for the augmentation ratio as defined in Eq. 5.2 and use them
to investigate the SGE as nitrous oxide washin plateaus and uptake reaches
maintenance levels. Using equations obtained in  4.3, we will first derive
augmentation ratios for a homogeneous lung i.e. one with no V Q mismatch
and then extend the derivation to a non-homogeneous lung with a log normal
distribution of V Q ratios.
5.2.1 HOMOGENEOUS LUNG

AR in blood phase
Using the symbols defined in Chapter 4, from Eqs. 4.9 and 4.10, it is trivial
to find that:
AR ( blood ) = (1 − F ) (1 − FI )
VSGhl/ FG / VSG hl = (5.3)
AR in gas phase
Substituting for VI from Eq. 4.3 in Eq. 4.2, we obtain an expression for FSG
hl
/ FG :
 (VA / Q ) FISG  1− F 
FSGhl / FG =    (5.4)
 
 (VA / Q) + λSG   1 − FI 
When no first gas is present, VI = VA , and hence the expression for FSG
hl
can be
easily derived from Eq. 5.4. Thus:
AR ( gas ) =
F hl / F hl
SG / FG SG
=(1 − F ) (1 − FI ) (5.5)
We can conclude that in a homogeneous lung with constant outflow, the
augmentation ratios in the gas phase and in the blood phase are both equal to

(1 − F ) (1 − FI ) so that the magnitude of the SGE is identical in both phases and
after substituting the solution for F , (obtained from Eq. 4.7), is given by:
(VA / Q ) + λ(1 − F v )
AR(gas) = AR(blood) = (5.6)
(VA / Q ) + λ(1 − FI )
5.2.2 NON-HOMOGENEOUS LUNG

AR in blood phase
From Eqs. 4.11 and 4.12, the augmentation ratio for blood is given by:
n Vj   1 − Fj 
∑ 
    
j =1  λSG + V j Q j   1 − FI 

AR ( blood ) = (5.7)
n  Vj 
∑    
j =1  λSG + V j Q j 
Applying Eq. 4.7 to the jth compartment and simplifying, we obtain the following
expression for AR(blood):
 n  Vj V0  
 ∑   
 λ ( FI − F v )   j =1  λ (1 − FI ) + Vj Q j  − 1
AR ( blood ) =
1+    (5.8)
 λSG − λ (1 − FI )   Vj V0 
n 
 ∑
 
 λ + V Q  
 j =1  SG j j  
Note here that with 70% N2O ( λ = 0.47 ) and FI = 0.7 , the expression
 λSG − λ (1 − FI )  is positive for all λSG > 0.141 which includes all the volatile
anaesthetic agents in current use.
Assuming now that these series are both applied using a log normal
distribution of V Q ratios with standard deviation given by =

σ σV − σ Q (see

 3.4), we replace each series with its equivalent Lebesgue-Stieltjes integral as
described in  3.12. Thus, we have 31:
 λ ( FI − F v )   I 
AR ( blood ) =
1+   − 1 (5.9)
 λSG − λ (1 − FI )   I SG 
where
1
∞
dx n  Vj V0 
I ∫ λ (1 − FI ) e ≡ ∑ 
   
( ) j =1  λ (1 − FI ) + V j Q j 
2
2π −∞
x 2 /2
+ V0 Q 0 e( x −σ ) /2
(5.10)
1
∞
dx n Vj V0 
I SG ∫−∞ λ e x2 /2 + V Q e( x−σ )2 /2 ≡ ∑   (5.11)
2π SG ( 0 0 )  
j =1  λSG + V j Q j 
Using the substitution FI * ( F v FI ) for F v , we can now determine
AR(blood) for a range of values of λSG , σ , F v FI (representing different stages
of N2O equilibration). However, Peyton and co-workers actually use the N2O
uptake, VN2O , as an input variable (68-70). This enables us to eliminate the term
( FI − F v ) , as described below. Using Eq. 4.8, the result derived by Colburn,

Evans and West (10) for the uptake of N2O in a homogeneous lung and summing
the uptake in each compartment to obtain the total uptake, we obtain the
following equation:
n λVj ( FI − F v )
VNnhlO = ∑ (5.12)
2
j =1 (
λ (1 − FI ) + V Qj j )
31
Note that we have used the Eqs. 3.1 and 3.2 here rather Eqs. 3.11 and 3.12 in forming the integrals
but as mentioned previously in  3.12, either set of equations will do.

which by analogy, produces the following expression for ( FI − F v ) if we
assume a log normal distribution of ventilation-perfusion ratios.
VNnhl
O
FI − F v =2 (5.13)
λV I 0
The final expression relating the SGE to the uptake of N2O and the degree of
ventilation-perfusion mismatch in a non-homogeneous lung is therefore:
 1   VNnhl  1 1 
AR ( blood ) =
1 +  2O
     −  (5.14)
 λSG + λ (1 − FI )  V 0   I SG I
We use WR the previously defined washin ratio as a variable on the x-axis. Note
that given the assumption that only the first gas, N2O contributes to changes in
gas volume, V=

I VA + VN O .
2
SGE in gas phase

Whereas the second gas is absorbed by the same volume of blood whether
or not N2O is present, this is not the case in the gas phase. Here the second gas
is contained in a smaller volume whose size depends on the rate of N2O uptake.
The equations describing this phase are therefore more complicated. Only the
final result is given here.
 1   VNnhl   λSG I SG − λ (1 − FI ) I 
AR ( gas ) = 1 +    
2O
  
 λSG + λ (1 − FI ) I   V0  1 − λSG I SG 
(5.15)
Note that Eqs. 5.14 and 5.15 may both be simplified to the following
approximate form:
1 + ψ (1 − WR )
AR = (5.16)

where ψ is a variable that differs for gas and blood and incorporates the
effect of σ . Perhaps the most useful form of Eq. 5.15 involves the substitution
of (VI − VA ) for VN2O . This gives rise to the following equation:
1 + k (VI − VA )
AR = (5.17)
When AR = 1, there is no second gas effect so when the SGE occurs, we can
write SGE = k (VI − VA ) . 32 This indicates how the second gas effect can be both
a function of the nitrous oxide uptake (VI − VA ) and the degree of ventilation-
perfusion mismatch (which appears in the term k along with the solubility of the
first and second gases and the inspired concentration of the first gas).
SIMULATING PERSISTENT SECOND GAS EFFECTS
Figure 5.1 shows the results for the blood and gas phases for desflurane
( λ = 0.42), isoflurane ( λ = 1.4), and diethyl ether ( λ = 12.1). The arrow in each
diagram indicates the direction in which σ is increasing. The x-axis shows the
proportion of nitrous oxide washin, and the y-axis shows the augmentation ratio.
The most consistent feature of these diagrams is the divergence of the lines from
the point (1,1), the point on the x-axis representing complete washin of the
nitrous oxide, at which time expired alveolar ventilation becomes equal to
inspired alveolar ventilation, so that according to Eq. 5.17, the SGE disappears
completely.
32
In the ANESTHESIOLOGY article, the term VA was included in the denominator. Since we are dealing
with the constant outflow case in this simulation, the equations for AR and SGE have been simplified
further by incorporating the constant expired ventilation VA into the variable k.

Figure 5.1. Augmentation ratios for desflurane, isoflurane and diethyl ether.
Second gas effect for several anaesthetic gases as a function of the proportion of washin
of nitrous oxide for different values of σ . The SGE is expressed as an augmentation ratio
(AR), defined as the partial pressure of the second gas in the presence of nitrous oxide,
divided by that which would have existed in the absence of nitrous oxide, with all other
factors kept constant. Upper three diagrams, SGE in the gas phase; lower three diagrams,
SGE in blood. In each case, the augmentation ratio is plotted on the y-axis as a function
of the washin ratio. The long arrow in each diagram indicates the direction of increasing
σ values from 0 to 2 in increments of 0.25 units. The line for σ = 0 is the same for each
SG. Reproduced with permission from ANESTHESIOLOGY and Wolters Kluver Health Inc.

Each augmentation ratio plot appears to be a straight line. The line for
σ = 0 , which represents a homogeneous lung, is the same for both gas phase and
blood phase, as expected for a “single compartment” lung with full equilibration
between gas and blood phases. In addition, it is the same for all three gases, i.e.
it is independent of the solubility of the second gas in blood.
The effect of increasing σ and hence the degree of V Q mismatch is seen
to be opposite for the blood and gas phases. In blood, an increase in σ increases
the augmentation ratio, while in the gas phase, the augmentation ratio decreases.
Moreover, in blood, this amplification is greatest for the least soluble gas,
desflurane, and least for the most soluble gas, diethyl ether, while in the gas
phase, the opposite is true. Results obtainable for sevoflurane ( λ = 0.67) are
similar to those shown for desflurane.
THE ROLE OF THE VARIABLE k
Equation 5.17 provides an explanation of how the SGE can continue, even
after nitrous oxide uptake decreases to levels seen in the “maintenance phase”.
Thus, it allows us to overcome the Severinghaus objection (81). The term
k (VI − VA ) is responsible for the SGE. Here the difference (VI − VA ) is produced
by nitrous oxide. A decrease in this difference as nitrous oxide uptake decreases
can be compensated for by a simultaneous increase in k, indicating that
“shrinkage” need not be the only factor involved in producing the SGE. We have
shown that k is affected by the degree of V Q mismatch and that the effect of
V Q mismatch on k differs in the gas phase and in blood. With no V Q
mismatch, σ = 0 and k = 1 in both phases. This is the case that corresponds to
the classical “volume shrinkage” explanation of the SGE. As V Q mismatch

increases, σ increases. In blood, as σ increases, k also increases and is greater
than 1. In the gas phase, as σ increases, k decreases and is less than 1. The term
k may be thought of as a scaling factor, controlling the degree to which the
“volume shrinkage”, (VI − VA ) , is expressed. Note however, that although we
have shown that it is possible for V Q mismatch to influence the SGE through
its contribution to k, this still does not explain exactly how this occurs. This
question is explored further in Chapter 7.
ADVANTAGES OF THE MATHEMATICAL MODEL
The mathematical model used allows contributions from the SG solubility
and V Q mismatch to be assessed separately. Kennedy (41) has made the
following comment in relation to this work: “A key component of the technique
is the investigation of the effect of varying V Q ratios across the lung. These
are very difficult to measure, let alone control, in a real-world experiment.”
RELATIONSHIP TO EXPERIMENTAL WORK
Our findings are consistent with experimental clinical findings of Peyton
et al. (71), that the magnitude of the SGE on arterial sevoflurane partial pressures
was significantly greater than that measured simultaneously on end-tidal partial
pressures in patients immediately after induction of anaesthesia and for 30
minutes afterward. However, these authors did not attribute the finding to a
concomitant diminution in SGE in the gas phase. Our model provides a more
precise explanation for their findings. The augmentation to blood desflurane
partial pressures at 20 to 30 minutes predicted by the model is of the order of 10
to 15% with moderate degrees of V Q scatter. This compares to a mean 12%
increase in arterial sevoflurane partial pressures at this time point in ventilated

patients measured by Peyton et al. The lesser augmentation in gas phase
desflurane concentrations predicted here is also broadly consistent with the
findings of Taheri and Eger (86), and those of Hendrickx et al. (34) for end-
expired sevoflurane concentrations in anaesthetized patients.
It is evident from Figure 5.1 that the SGE in blood and its augmentation
by increasing V Q scatter is greater for the least soluble second gas, desflurane.
This finding is remarkable, as lung modelling in the absence of nitrous oxide
predicts that increasing V Q scatter, with a reduction in effective alveolar
ventilation, impairs alveolar-capillary gas exchange more severely for less
soluble gases (10), a prediction recently confirmed in an animal study of
desflurane and isoflurane anaesthesia using methacholine inhalation to induce
increased V Q inhomogeneity (49). This would suggest that the addition of
nitrous oxide to the inspired gas mixture may maintain the efficiency of the lung
in exchanging less soluble modern volatile agents in the face of V Q scatter
induced by anaesthesia itself.

CLINICAL IMPLICATIONS
While the SGE in the gas phase is most easily detected using end-tidal gas
concentration monitoring by the infrared analyzers available in most operating
theatres today, augmentation in the blood phase is not normally measured but is
important because it most directly affects anaesthetic partial pressures in blood
and therefore depth of anaesthesia. An obvious clinical implication of these
findings is that calculations of the minimum alveolar concentration 33 (MAC)
based on end-tidal concentration measurements may well underestimate the
depth of anaesthesia when nitrous oxide is supplemented with a volatile agent.
This represents a new factor that should be taken into consideration in addition
to other relevant factors (3, 23, 37, 61), when interpreting a MAC reading during
anaesthesia.
33
This is defined as the concentration of the anaesthetic agent in 100% O2, that prevents movement in
response to skin incision in 50% of unpremedicated subjects at sea level (1 atm). Prior to the invention
of the BIS and Entropy depth-of-anaesthesia monitors, this was the commonly used measure of depth
of anaesthesia. The more potent the anaesthetic agent, the lower the MAC value. This measure of
anaesthetic depth was introduced into clinical anaesthetic practice by Merkel & Eger (60).

CONCLUSION
In this chapter we have applied the equations derived in Chapter 4 to
simulate the second gas effect as nitrous oxide uptake falls to the low levels seen
during maintenance anaesthesia. The technique allows the effects of SG
solubility and V Q mismatch to be assessed separately. It is shown that the
increase in V Q mismatch that is known to occur during anaesthesia can
adequately explain the persistent SGE observed experimentally. It is also shown
that as the SGE increases with V Q mismatch in blood, it is simultaneously
diminished in the gas phase.
An inevitable conclusion of relevance clinically is that MAC readings,
which are based on end-tidal gas analysis, may underestimate the depth of
anaesthesia when volatile agents are used together with nitrous oxide.

: Partial pressure-solubility diagrams
In the last chapter we verified Peyton’s claim that the second gas effect
can persist into the maintenance phase of nitrous oxide anaesthesia and
confirmed that this could be achieved through the increase in ventilation-
perfusion mismatch known to occur during anaesthesia. We now turn our
attention to a search for other manifestations of this strange interaction between
V Q mismatch, low solubility and shrinkage of gas volume. Because this
combination has not appeared on the radar of Respiratory Physiologists in the
past, it seems appropriate to revisit some of the previous results which might be
affected. This is the subject of the current chapter.
THE RACE FOR MIGET
During the 1970s, John West’s group in San Diego was engaged in a race
with Hermann Rahn’s group in Buffalo to develop a practical method of
obtaining actual distributions of V Q ratios in the lung. Their work resulted in
the Multiple Inert Gas Elimination Technique or MIGET (88), a test which is
available today in various specialized centres around the world. West recruited
a professional mathematician, John Evans to work in his team. Evans was a
qualified medical practitioner who had abandoned medicine for a career in
mathematics. Because of his medical training, he understood clearly the
problems which confronted the team and possessed the mathematical skills to
help solve them. A series of publications followed in which various aspects of
gas exchange were formalized mathematically in a way not previously attempted
Chapter 6: Partial pressure-solubility diagrams 79

(25). We have already used some of the results of this collaboration in previous
chapters.
In one of these studies, West et al. (89) obtained partial pressure-solubility
diagrams in the case of a single soluble inert gas present in the inspired gas
mixture together with other gases not undergoing net transfer 34. Their study
focused on the elimination from the lung of a gas supplied at constant partial
pressure in mixed venous blood. Colburn et al. (10) explored further the
exchange of inert gases in such a system. They concluded that mismatching of
ventilation and blood flow must reduce the pulmonary uptake of any gas with a
linear dissociation curve i.e. any gas that obeys Henry’s Law. This conclusion
was stated explicitly by Evans et al. (26), in a separate publication. Because of
their complex interaction with haemoglobin, neither oxygen nor carbon dioxide
obeys Henry’s Law. Evans et al. did investigate situations in which oxygen and
carbon dioxide might not obey this rule, but these were extremely unlikely to be
met under normal conditions. Using a computer model which generated a wide
variety of V Q distributions, Neufeld et al. also concluded that inert gas a-A
differences are a direct reflection of V Q distribution (64).
These findings underpin the generally accepted principle that an increase
in V Q mismatch must always lead to a reduction in gas transfer in the lung, a
principle easily understood from Figure 3.4. However, these authors did not
consider possible effects associated with net gas volume changes during gas
exchange. In this chapter we use N2O uptake as a tool to investigate the effect of
34
As explained on page 36, this is equivalent to assuming that the soluble gas is accompanied by
an insoluble filler gas.
80 Chapter 6: Partial pressure-solubility diagrams

net gas volume changes on the partial pressure-solubility diagram and its effect
on gas uptake in non-homogeneous lungs.
EQUATIONS
We begin once more by presenting the relevant equations.
6.2.1 STEADY-STATE EQUATIONS IN A HOMOGENEOUS LUNG

Consider gases with linear dissociation curves, commonly referred to as
inert gases, being exchanged under steady-state conditions. If such a gas is
delivered in the inspired gas mixture at a rate VI FI and removed at a rate VA FA
in a homogeneous lung of given VA Q t , then the uptake of the gas is
VI FI − VA FA . This must equal the uptake by blood which is given by
λQ t ( Fa − F v ) , where Fa and F v are equal to Pa ( PB − PH 2O ) and
Pv (P B )
− PH 2O respectively, so that we have the following equation (analogous
to Eq. 4.2):
VIPI − VAPA= λQt ( Pa − P v ) (6.1)
For the diluent gas, which is treated as if it were completely insoluble, we
have:
VI (1 − FI ) = VA (1 − FA ) (6.2)
Assuming complete equilibration between gas and blood so that Pa = PA ,
these equations give rise to the following general equation for the alveolar and
arterial partial pressures of the inert gas (analogous to Eq. 4.7) :
P=
a P=
A
( VA Q ) PI + λ (1 − FI ) Pv (6.3)
( VA Q ) + λ (1 − FI )

Given that VI Q = (VI VA ) . (VA Q ) , Eq. 6.2 provides the following
expression for VI Q :
( VA Q ) + λ (1 − F v ) VA Q
(
VI Q = ) VA Q + λ (1 − FI ) ( ) (6.4)
( )
6.2.2 INSPIRED VENTILATION WITH LOG NORMAL DISTRIBUTION
In a lung model consisting of N parallel compartments, the inspired
ventilation is the sum of the individual values for each compartment:
N
VI = ∑ VI j (6.5)
j =1
We have previously shown (see  3.10) that with a log normal distribution
of V Q ratios with standard deviation σ , Eq. 6.5 may be expressed as the
equivalent Lebesgue-Stieltjes integral 35 as N → ∞ and becomes:
 λ ( FI − F v ) 
VA
{e }  
+∞
=VI ∫
− x 2 /2
+ dx (6.6)
( )
2
 VA Q e + λ (1 − FI ) e 
2π −∞ − ( x − σ ) /2 x 2
/2
The left-hand term in the integrand produces the log normal distribution of
expired ventilation-perfusion ratios, the right-hand term gives the uptake of
nitrous oxide. Once VA , Q , σ , and F v are specified, Eq. 6.6 provides a one-to-
one relationship between VI and FI , and hence allows the calculation of the
inspired concentration of nitrous oxide associated with the desired total inspired
V
alveolar ventilation. Letting Δ= (VI − VA ) we may now determine the required
35
Eq. 6.6 may take several alternative forms if the variable x is substituted as discussed in 3.8.
These are equivalent to each other. As mentioned previously in 3.12, we choose the most
suitable form for use in each particular situation.

values of ΔV which we use as an independent variable for investigation, along
with σ , the degree of mismatch between ventilation and blood flow. We use the
terms “volume shrinkage”, “volume contraction” and “net gas volume uptake”
interchangeably, when referring to ΔV . For convenience, we set F v in Eq. 6.6
equal to 0. Alveolar ventilation VA is set at 4 l/min and pulmonary blood flow
Q t is set at 5 l/min. For individual figures, values of ΔV were chosen so as to
demonstrate differences when these occur, while also lying within the range
believed to occur clinically during N2O anaesthesia.
6.2.3 LEBESGUE-STIELTJES INTEGRALS FOR THE FIRST GAS

For the lung model consisting of N parallel compartments, each having a
different ventilation-perfusion ratio, the partial pressure of the mixed expired
alveolar gas PA is the ventilation-weighted mean of the individual
compartmental partial pressures:
∑VA PA
j =1
j j
PA = N
(6.7)
∑VA
j =1
j
Setting P v = 0 in Eq. 6.3 we have ,
P AFG
=
VA Q t (
I1
) (6.8)
PI FG 2π
∞
dx
where I1 = ∫ ( ) +   x
(6.9)
λ (1 − FI ) e (VA Qt ) e
x 2 + 2 σx − σ 2 /2 2
/2
−∞

The partial pressure of the mixed pulmonary end-capillary blood P a is the
perfusion-weighted mean of the individual compartmental partial pressures:
∑ Q PA
j =1
j j
Pa = N
(6.10)
∑ Q
j =1
j
from which we obtain the following expression:
P aSG
=
(
VA Q t
I3
) (6.11)
PI SG 2π
∞
dx
where I3 = ∫ λ (1 − FI ) e (6.12)
( )
2
+ VA Q t e( x −σ )
2 /2
x /2
−∞
6.2.4 LEBESGUE-STIELTJES INTEGRALS FOR THE SECOND GAS

For further treatment of the SGE, we again use the subscript “SG” to
distinguish the second gas from the first gas, nitrous oxide, whose abbreviations
are left, for simplicity, with no subscript.
From Eq. 6.3, and setting P v = 0 for the second gas we have:
FjSG  Vj Q j   
=  (VI VA )j (6.13)
FI SG  λSG + Vj Q j 
where (VI VA ) is the ratio of the inspired ventilation to the expired ventilation
j
in the jth compartment. This ratio may be obtained from Eq. 6.4 and results in the
following equation:
FjSG  Vj Q j   λ + Vj Q j 

=    (6.14)
FI SG  λSG + Vj Q j   λ (1 − FI ) + Vj Q j 

λ′ λ (1 − FI ) , and expressing Eq. 6.14 using partial fractions:
Setting =
=
FjSG ( λ − λSG ) (Vj Q j) −
( λ − λ' ) (Vj Q j) (6.15)
FI SG ( λ'− λSG ) ( λSG + Vj Q j) ( λ'− λSG ) ( λ' + Vj Q j)
Applying Eqs. 6.7 and 6.10, and converting sums to Lebesgue-Stieltjes
integrals we obtain:
P ASG (VA Q t )
= {( λ − λ ) I − ( λ − λ') I } (6.16)
( λ' − λSG )
SG 2 1
PI SG 2π
∞
dx
where I2 = ∫ (x ) /2 + V  x
(6.17)
( A Qt ) e
2 2
+ 2 σx − σ 2
−∞ λSGe /2
P aSG (
VA Q t )
and =− {( λ − λSG ) I 4 − ( λ − λ') I3 } (6.18)
PI SG ( λ' − λSG ) 2π
∞
dx
where I4 = ∫λ (6.19)
( )
2
e x /2 + VA Q t e( x −σ ) /2
2
−∞ SG
6.2.5 INSPIRED N2O CONCENTRATION FOR ZERO A-a GRADIENT

When seeking the value of FI N2O that results in P ASG and P aSG being equal
(see  6.7), the following equation is solved for λ′ in MATLAB:
PASG P aSG
− 0
= (6.20)
PI SG PI SG
and the result converted to a value for FI N2O by inverting the relationship
between λ′ and FI N2O :
λ'
FI N2O = 1 − (6.21)
λ

THE PARTIAL PRESSURE-SOLUBILITY DIAGRAM
Solving Eqs. 6.16 and 6.18, graphs may now be drawn for P A PI and
P a PI as a function of λ for different values of σ and ΔV .
6.3.1 THE CLASSICAL DIAGRAM

We begin by drawing a partial pressure-solubility diagram of the form
described by West, Wagner and Derks (89). This is achieved in our model by
setting the inspired nitrous oxide concentration to zero, so that with the very low
inspired concentrations of the other soluble gas present, there is effectively no
change in net gas volume. The resulting curves are shown in Figure 6.1A.
When there is no V Q mismatch, i.e. σ = 0, alveolar and arterial partial
pressures are equal. This is indicated by the black stippled curve. Moving along
this curve towards the left end of the x-axis, we are dealing with less soluble
gases – very little is transferred to blood so that the alveolar (and hence arterial)
partial pressure approaches the partial pressure in the inspired gas mixture and
both P A PI and P a PI approach 1.
Figure 6.1. Standard partial pressure-solubility diagrams.

Alveolar and arterial partial pressures as a fraction of the inspired partial pressure vs
solubility expressed as log10 ( λ ) . A: P A PI ( ) and P a PI
( ) with no net volume change i.e. ΔV = 0 and σ = 1.5 The stippled black curve
( ) represents the situation with σ = 0 in which case P A PI and P a PI are
equal; B: no ventilation-perfusion mismatch i.e. σ = 0, P A PI and P a PI are equal,
lower curve ΔV = 0 ( ) and upper curve ΔV = 600 ml/min ( ); C: the
curves in A and B are summed for gas ( ) and blood ( ). In each case, the
stippled green vertical line on the left indicates the position of sulphur hexafluoride, the
stippled green vertical line on the right indicates the position of methoxyflurane.

As we move along the black stippled curve towards the right end of the x-
axis, we are dealing with more soluble gases – more and more of the alveolar
gas is transferred to blood so that the alveolar (and hence the arterial) partial
pressure approaches zero, and so do the respective ratios P A PI and P a PI .
The presence of V Q mismatch acts as an impediment to gas exchange –
less gas is removed from the alveolar gas mixture producing a higher partial
pressure in the gas phase and a lower partial pressure in blood. Thus, the curves
in Figure 6.1A for alveolar gas and arterial blood separate, producing an
alveolar-arterial partial pressure difference. At each end of the solubility scale,
P A PI and P a PI behave as before and approach 1 and 0 respectively. This is
the classical description given by West, Wagner and Derks (89).
Figure 6.1B depicts partial pressure-solubility diagrams in the absence of
V Q mismatch i.e. with σ = 0. The black stippled curve is the same as in Figure
6.1A and represents the situation where there is no net gas volume change
( ΔV = 0 ) ; the solid black line represents the situation with a net gas volume loss
of 600 ml/min ( ΔV = 600 ) . The difference between the two curves reflects the
magnitude of the second gas effect. The effect is greater with gases of low
solubility, a fact noted by Stoelting and Eger (84). Until the work of Peyton et
al. (68-70), no particular effect was ascribed to the addition of V Q mismatch.
If the effects were merely additive, they would produce the result shown in
Figure 6.1C.

This page has deliberately been left blank.

6.3.2 THE EFFECT OF NET GAS VOLUME UPTAKE
Figure 6.2 shows what actually happens in our model when a net gas
volume reduction is combined with V Q mismatch. It shows the effect of
increasing ΔV from 0 ml/min to 600 ml/min on the alveolar and arterial partial
pressures in a homogeneous lung ( σ = 0) and a lung with significant V Q
mismatch ( σ = 1.5).
With no volume change ( ΔV = 0 ) , the graph of P A PI is displaced to the
right by V Q scatter, as in Figure 6.1A. A similar displacement occurs when
ΔV = 600 ml/min. Now it is the increase in P A PI of the continuous blue curve
relative to the continuous black curve that reflects the impaired uptake in the
presence of an increased degree of mismatch. For arterial blood, when ΔV = 0 ,
the curve of P a PI is displaced to the left by V Q scatter as in Figure 6.1A.
With ΔV = 600 , the graph of P a PI with σ = 1.5 (continuous red curve)
greatly exceeds that for σ = 0 (continuous black curve) for much of its course,
especially at low solubilities ( λ < 0.05 ) . This reflects the increased uptake which
occurs with significant gas volume shrinkage. The effect of the contraction in
net gas volume on P A PI and P a PI at each value of σ is obtained by
comparing the stippled curve with the continuous curve for each colour. The
greater magnitude of the SGE in blood than in the gas phase is indicated by the
larger area between the two red curves in Figure 6.2B when compared with the
area between the two blue curves in Figure 6.2A.

Figure 6.2. Partial pressure-solubility diagrams corrected for volume change.
Alveolar and arterial partial pressures as a fraction of the inspired partial pressure vs
solubility expressed as log10 ( λ ) . A: P A PI with σ = 0, ΔV = 0 ( ); σ = 0,
ΔV = 600 ml/min (( ) σ = 1.5, ΔV = 0 (

); ); σ = 1.5, ΔV = 600 ml/min
( ) and B: P a PI with σ = 0, ΔV = 0 ( ); σ = 0, ΔV = 600 ml/min
( ); σ = 1.5, ΔV = 0 ( ); σ = 1.5, ΔV = 600 ml/min ( ). The

stippled green vertical line on the left indicates the position of sulphur hexafluoride, the
stippled green vertical line on the right indicates the position of methoxyflurane.

THE AUGMENTATION-SOLUBILITY RELATIONSHIP
A useful measure of the magnitude of the SGE is the augmentation ratio
(AR) which was defined in  5.2 as the partial pressure of the second gas in the
relevant phase (gas or blood) in the presence of N2O divided by the partial
pressure the second gas would have in that phase in the absence of N2O, all other
factors being held constant (24, 45). Augmentation ratios may be obtained for
the gas phase and for blood by dividing P A PI and P a PI at any given value
of ΔV with their equivalent values when ΔV = 0 . In each case, all other
variables must be held constant. AR(gas) and AR(blood) may then be graphed
as functions of λ .
The augmentation or fractional increase in partial pressure of a gas due to
gas volume shrinkage may be derived in Figure 6.2 by dividing the value of
P A PI or P a PI when ΔV = 600 by its value when ΔV = 0 for the value of
λ associated with that gas. In Figure 6.3A we have shown the augmentation ratio
for gas (blue) and blood (red) for ΔV = 200 ml/min (stippled curve) and 600
ml/min (continuous curve) with σ = 1.5; augmentation, (and hence the SGE) is
greater at the higher value of ΔV , an effect most noticeable as the solubility of
the gas is decreased.
Figure 6.3B shows the augmentation ratio for gas and blood for a range of
σ values with ΔV set at 100 ml/min. In the absence of V Q mismatch, the
augmentation ratio is 1.2 (horizontal black line). Increasing the degree of
mismatch has opposite effects on the gas phase (blue) and blood (red). As σ
increases, the SGE increases in blood, but decreases in the gas phase (black
arrows). This in line with our previous findings (45) as shown in Figure 5.1.

Figure 6.3. Augmentation-solubility diagrams.
A: Augmentation-solubility diagram with σ = 1.5. Gas phase, ΔV = 200 ml/min
( ); ΔV = 600 ml/min ( ). Blood, ΔV = 200 ml/min
( ); ΔV = 600 l/min ( ). B: effect of increasing σ with ΔV constant and

set at 800 ml/min. The arrows indicate the direction of increase of ventilation-perfusion
mismatch ( σ = 0, 0.5, 1.0, 1.5 and 2.0).

ALVEOLAR-ARTERIAL PARTIAL PRESSURE DIFFERENCES
An estimate of the A-a difference may be obtained by multiplying the
difference { P A PI − P a PI } by the partial pressure of the second gas in the dry
inspired gas mixture. For example, when 1% sevoflurane is present in the dry
inspired mixture it is necessary to multiply the difference by 7.13 to obtain an
estimate of the A-a difference in mm Hg.
The estimate may be inferred from the partial pressure-solubility diagram.
For example, in Figure 6.4A-D which follow, P A PI and P a PI have been
drawn with σ = 0 (stippled black line) and σ = 1.5 (continuous lines, blue = gas,
red = blood) for increasing values of ΔV (from 0 to 800 ml/min).
Figure 6.4. More partial pressure-solubility diagrams.

Alveolar (blue) and arterial (red) partial pressures as a fraction of the inspired partial
pressure (on the y-axis) vs log10 ( λ ) (on the x-axis) for a range of λ values with
(clockwise) σ = 1.5, A: ΔV = 0, B: ΔV = 200, C: ΔV = 400, and D: ΔV = 800
ml/min. The stippled black curve in each diagram ( ) is associated with no
mismatch of ventilation and perfusion i.e. σ = 0.

When σ = 0, there is no mismatching between ventilation and blood flow,
P A PI and P a PI are equal. This is represented by the single black stippled
curve in each figure regardless of the value of ΔV . Notice that in each case, the
blue line representing P A PI always lies above or coincides with the black line.
However, while the red line representing P a PI lies below the blue line for
P A PI on the right of the solubility axis, it crosses the blue line as one moves
left along the solubility axis. This represents the situation where the partial
pressure in mixed pulmonary end-capillary blood exceeds that in mixed alveolar
gas. The point of intersection is affected by the shrinkage factor and moves to
the right as ΔV is increased.
Figure 6.5. Alveolar-arterial partial pressure difference for several gases.

Graphs of the A-a partial pressure difference (mm Hg) for (clockwise) diethyl ether (A),
sevoflurane (B), ethylene (C) and sulphur hexafluoride (D) as a function of σ (the degree
of V Q mismatch) for ΔV = 0, 200, 400, 600, 800 and 1000 ml/min. In each case, the
direction of increase in ΔV is shown by the red arrow. Note the change in the range of
values shown on the y-axis for sulphur hexafluoride.

The effect of increasing ΔV is shown in Figure 6.5 for several gases, in
order of decreasing solubility: diethyl ether, sevoflurane, ethylene, sulphur
hexafluoride 36 ( λ = 12, 0.59, 0.14, 0.0076, respectively). It is assumed that the
inspired concentration of each gas is 1 percent, which represents an inspired
partial pressure of 7.13 mm Hg.
The red arrow in each diagram shows the direction of the increase in ΔV
which is varied from 0 to 1 l/min in increments of 200 ml/min. Ether exhibits a
worsening in A-a gradient as ΔV is increased, but this is almost imperceptible.
In the case of sevoflurane and ethylene, the A-a partial pressure gradient is
reduced and moves towards zero, the value seen when σ = 0 and there is no
ventilation-perfusion mismatch. With sulphur hexafluoride, the A-a partial
pressure difference is less than zero when ΔV is 400 ml/min or greater, with the
partial pressure in pulmonary end-capillary blood predicted to be 1 mm Hg
higher than that in mixed alveolar gas, when σ = 1.2 and ΔV = 600 ml/min.
36
We use both the Australian spelling “sulphur hexafluoride” and the US spelling “sulfur
hexafluoride” in this thesis.

Figure 6.6. Uptake-solubility diagrams.
Graph of R( λ ) vs log10 ( λ ) for a range of λ values. A: in the absence of net volume
change. The red arrows indicate the direction of increase of σ from 0 to 4 in increments
of 0.5; B: with a net volume change of 1000 ml/min (BTPD), so that ΔV VA = 0.25. Red
arrows as in A. The blue arrow indicates the point where the uptake ratio of sulphur
hexafluoride exceeds 1; C: with the degree of V Q mismatch fixed at σ = 2, the net
volume change has been increased in the direction of the red arrows from 0 ml/min to
1000 ml/min (BTPD), in increments of 200 ml/min. In each diagram, the solid vertical
green line is drawn at log10 ( 0.8 ) ; the position of sulphur hexafluoride is shown by the
stippled green vertical line on the left, that of methoxyflurane by that on the right.

THE UPTAKE-SOLUBILITY RELATIONSHIP
As an index of gas exchange, Colburn, Evans and West studied R(λ), the
ratio of the gas exchange rate of a compartmental model over the exchange rate
of the homogeneous model with the same total expired alveolar ventilation and
blood flow (10). In their model, R(λ) is always less than 1, indicating that gas
exchange in a non-homogeneous lung is always less than in the equivalent
homogeneous lung. We use their expression for R(λ) and apply it to an inert gas,
present in the inspired gas mixture in low concentrations in the presence of net
volume changes generated by N2O uptake.
Many of the above results are consistent with gas uptake in the presence
of ventilation-perfusion mismatch exceeding that in the absence of ventilation-
perfusion mismatch, a possibility that is the subject of our next investigation. In
Figure 6.6, we compare the situation which applies when V Q mismatch occurs
in the absence of net gas volume changes with two scenarios in the presence of
such changes.
Figure 6.6A shows a series of such graphs for values of σ ranging from 0
to 4 in increments of 0.5. With very low inspired concentrations of inert gas there
is no net volume change during gas exchange and the family of curves is
symmetric about the vertical green line drawn at x = log10 (VA Q ) . As σ is
increased, R(λ) falls as predicted by Colburn et al. (10) and is never greater than
1.
This is the classical result described by Colburn et al. (10), with which we
now compare our two scenarios. Figure 6.6B shows the effect of increasing the
value of σ from zero with the net volume uptake fixed at 1000 ml/min (BTPD),

while Figure 6.6C shows the effect of increasing the value of ΔV from zero with
the degree of mismatch fixed at σ = 2. In each case, once ΔV is greater than
zero, the curves are no longer symmetric about the vertical line x = log10 (VA Q )
; the minimum for each curve has been shifted to the right; each curve begins
above the R(λ) = 1 line and ends below it. At the left end of each curve in Figure
6.6B, R(λ) increases as σ increases, while at the right end, R(λ) decreases as σ
increases. At the left end of each curve in Figure 6.6C, R(λ) increases as ΔV
increases, while at the right end, R(λ) decreases as ΔV increases.
Where R(λ) is greater than 1, uptake of a low inspired concentration of an
inert gas is predicted to be greater in a non-homogeneous lung than in the
equivalent homogeneous lung. However, only sulphur hexafluoride appears to
exhibit this property (see blue arrow in Figure 6.6B) and only then under extreme
conditions, which are probably not attainable clinically ( ΔV = 1000 ml/min, σ
= 4).
PARTIAL PRESSURE REVERSAL DURING GAS UPTAKE
According to West, Wagner and Derks, during gas uptake, the alveolar
component of the A-a difference is equal to λ over V Q times the arterial
component and must therefore always be positive (89). However, this rule
clearly cannot apply at the lower end of the solubility scale in Figure 6.4B-D.
Our findings are shown in Figure 6.5. With VA = VI (i.e. ΔV = 0), the A-a partial
pressure difference always increases as the degree of V Q mismatch increases
but less so as the solubility of the gas is decreased. As the difference between
VA and VI is increased with the less soluble gases shown in Figure 6.5B-D, the

A-a difference decreases. In the case of sulphur hexafluoride, the least soluble
of the gases, the pressure gradient is reversed, with the partial pressure in blood
predicted to be higher than that in the gas phase, even though the movement of
the sulphur hexafluoride is in the direction from gas phase to blood. One would
normally expect the partial pressure of a gas to be lower at all points downstream
during gas uptake, for example as is commonly described in the transfer of
oxygen from inspired air to the mitochondrion - the so-called “oxygen cascade”
(5).
Whether this phenomenon occurs in a clinical setting will depend on the
values of λ, σ , and (VI − VA ) , the factors controlling augmentation in Eq. 5.17.
While an increase in the degree of V Q mismatch favours augmentation through
its effect on the term k in Eq. 5.17, it impedes nitrous oxide uptake and therefore
can reduce augmentation through its effect on the “shrinkage” term, (VI − VA ) .
Increasing the inspired nitrous oxide concentration can compensate for this but
is limited by the maximum concentration compatible with safe anaesthesia.
These competing effects are shown in Figure 6.7. Each point on the blue
coloured curved surface in Figure 6.7 represents a point at which the alveolar
and arterial partial pressures are equal. This surface is shown as a function of the
gas solubility (x-axis) and the degree of V Q mismatch (y-axis). On the z-axis
is shown the FI N2O corresponding to each point on the surface. The surface
slopes upwards and backwards from the bottom right to the top left. A yellow
horizontal plane is drawn for FI N2O = 0.7, the highest inspired concentration

Figure 6.7. Alveolar-arterial partial pressure reversal.
Three-dimensional surface showing all points for which the alveolar and arterial partial
pressures of the second gas are equal, shown as a function of the inspired nitrous oxide
concentration necessary to produce the required volume change for the specified values
of λ and σ . The surface is coloured blue and slopes upwards and backwards from the
bottom right (dark blue) to the top left (light blue). The yellow horizontal plane indicates
an FI N O = 0.7 and intersects the blue surface as shown. The green arrow points to the
2
region between the yellow horizontal plane and the dark-blue curved surface compatible
with the phenomenon of A-a partial pressure reversal. When σ = 0, P A = P a for all
values of FI N O so the graph has been offset slightly from σ = 0 to facilitate discussion.
2
which can be used safely in most healthy subjects without producing hypoxia
(55). In the region beneath the blue surface (red arrow), the inspired nitrous oxide
concentration is too low to produce the required volume change. The region
above the yellow plane on which FI N2O = 0.7 (black arrow) is associated with
an “unsafe” FI O2 . In the region above the dark blue surface and below the yellow
FI N2O = 0.7 plane (green arrow), it is possible for nitrous oxide uptake to
generate a reversal of the alveolar-arterial partial pressure gradient for the second

gas. During anaesthesia σ generally lies between 0.75 and 1.5. With σ = 0.75,
the highest solubility for which A-a reversal can occur is approximately 0.35 so
that none of the volatile agents in common use can be expected to demonstrate
A-a partial pressure reversal during uptake. However at least 6 other gases may
demonstrate the effect: xenon ( λ = 0.2), ethylene ( λ = 0.14), ethane ( λ = 0.092),
methane ( λ = 0.038), nitrogen ( λ = 0.014) and sulphur hexafluoride ( λ =
0.0076). With σ = 1.5, the highest solubility for which A-a reversal can occur is
approximately 0.15 so that only ethane, methane, nitrogen and sulphur
hexafluoride can be expected to demonstrate the phenomenon. Note that
although A-a partial pressure reversal is theoretically possible, the force driving
gas uptake is actually the inspired-mixed venous partial pressure gradient.
ALVEOLAR-ARTERIAL PARTIAL PRESSURE GRADIENT FOR N2
So far, we have only considered a second gas with a mixed venous partial
pressure of zero, undergoing net uptake in the presence of volume “shrinkage”
produced by nitrous oxide uptake. A slightly different scenario was investigated
by Canfield and Rahn, who studied alveolar-arterial nitrogen partial pressure
differences in anaesthetized dogs being ventilated with air (8). Here the
difference between inspired and expired alveolar ventilation is smaller than
during nitrous oxide anaesthesia. It is produced by the discrepancy between
oxygen uptake and carbon dioxide output but there is no net exchange of
nitrogen. They showed the mean arterial pressure of nitrogen was greater than
that of mean alveolar gas and ascribed it to the variation in V Q ratios
throughout the lung. The situation may be mimicked in our model by setting P v
equal to P a for the inert gas involved as described in the next section.

6.8.1 LEBESGUE-STIELTJES INTEGRALS WITH P v > 0
We can incorporate this scenario in our model by including the mixed
venous partial pressure of N2 in the mass balance and setting it equal to the
arterial partial pressure. P v is then an unknown to be determined for N2. It is
still necessary to determine PI for N2O uptake to produce the desired value of
ΔV . After equilibration in the jth compartment, for the second gas we now have
from Eq. 6.3, the following modified form of Eq. 6.14:
 PASG   Vj Q j   λ + Vj Q j  λSG ( P v PI ) SG

=  + (6.22)
         
 PI SG  j  λSG + Vj Q j   λ (1 − FI ) + Vj Q j  λ + Vj Q j
SG
P ASG N
Pj
So, since = ∑ SG (Vj VA ) , the mean alveolar partial pressure is given
PI SG j =1 PI SG
by:
P ASG N
  Vj Q j   λ + Vj Q j  λSG ( P v PI ) SG 
∑ (Vj VA )  λ       
+ 
PI SG j =1  SG + Vj Q j   λ (1 − FI ) + Vj Q j  λSG + Vj Q j 
(6.23)
P aSG N
Pj
and since ( )
= ∑ SG Q j Q t , the mean pulmonary end-capillary blood
PI SG j =1 PI SG
partial pressure is given by:
P aSG N
  Vj Q j   λ + Vj Q j  λSG ( P v PI ) SG 
∑ ( Q j Q t )  λ       
+ 
PI SG j =1  SG + Vj Q j   λ (1 − FI ) + Vj Q j  λSG + Vj Q j 
(6.24)
Defining I5 as follows:
∞
dx
I5 = ∫  e( x )
(6.25)
( )
2
2 − 4 σx + 2 σ 2 /2
−∞ λSG e(
x − σ ) /2
+ VA Qt

Eqs. 6.23 and 6.24 may be converted to the following equations containing
λ′ λ (1 − FI ) , as defined previously:
Lebesgue-Stieltjes integrals in which =
P ASG (VA Qt ) λSG ( P v PI ) SG

= {( λ − λ ) I − ( λ − λ') I } + I 4 (6.26)
( λ' − λSG )
SG 2 1
PI SG 2π 2π
P aSG (VA Qt ) λSG ( P v PI ) SG

= {( λ − λ ) I − ( λ − λ') I } + I 5 (6.27)
( λ' − λSG )
SG 4 3
PI SG 2π 2π
Since P vSG = P aSG , we can eliminate the term containing P v from Eq. 6.27 and
substitute for ( P v PI ) SG in Eq. 6.26.
P aSG (VA Qt )

= {( λ − λ ) I − ( λ − λ') I } (6.28)
( λ' − λSG ) ( ) SG 4 3
PI SG 2π − λSG I 5
Note that if σ = 0, I1 - I 5 become:
2π
I= I= (6.29)
1 3
( 
λ'+ VA Qt )
2π
I= I= I= (6.30)
2 4 5
( 
λSG + VA Qt )
So, Eq. 6.28 becomes:
P aSG
=
(
λ + VA 
Qt ) (6.31)
PI SG (
λ' + VA 
Qt )

and Eq. 6.26 produces the same result so that if FI N2O > 0, λ′ < λ , alveolar and
arterial pressures are equal and both greater than the inspired partial pressure. If
λ ' = λ , alveolar, arterial and inspired partial pressures are all equal.
The results are shown in Figure 6.8. This graph may be compared with the
graphs in Figure 6.5. The red arrow shows the direction of increase in ΔV . This
is associated with an increase in the size of the a-A partial pressure gradient.
Canfield and Rahn obtained an average gradient of 16 mm Hg, indicated by the
horizontal green line in Figure 6.8. The curve for ΔV = 50 ml/min intersects this
line with σ approximately equal to 1.8. This is close to the upper limit of σ
values known to occur in humans during anaesthesia.
Figure 6.8. A-a partial pressure difference for nitrogen when R < 1.
Alveolar-arterial partial pressure difference for nitrogen as a function of σ , the degree of
ventilation-perfusion mismatch, for ΔV = 0, 25, 50, 75, 100 ml/min. These values of
ΔV are associated with a Respiratory Exchange ratio less than or equal to 1. The
direction of increase in ΔV is shown by the red arrow.

The situation that applies when the net volume movement is from blood to
alveolar gas is shown in Figure 6.9. Here we have set PI to zero for N2O and
determined the value of P v associated with the specified value of ΔV . This
gives rise to the following set of equations for the second gas, nitrogen:
λN2O F vN2O VA

ΔV = − Iv
2π
∞ (6.32)
dx
Iv = ∫
( )
2
+ VA Q t e( x −σ ) /2
2
x /2
−∞ λN O e
2
2π ΔV
F vN2O = − (6.33)
λN2O VA I v
PASG (VA Q t ) λSG ( P v PI ) SG

=
PI SG ( λSG − λ ) 2π
{( λ − λ ) H − ( λ − λ ) H } +
*
1
*
SG 2
2π
H3
(6.34)
P aSG (VA Q t ) λSG ( P v PI ) SG

=
PI SG ( λSG − λ ) 2π
{( λ − λ ) H − ( λ − λ ) H } +
*
4
*
SG 5
2π
H6
(6.35)
∞
dx
H1 = ∫ λe
−∞
( x 2 + 2 σx − σ 2 )/2
(
+ VA Q t e x /2
2
)
(6.36)
∞
dx
H2 = ∫λ
−∞ e ( x 2 + 2 σx − σ 2 )/2
(
+ VA Q t e x /2
2
)
(6.37)
SG
∞
dx
H3 = ∫λ
−∞ e x 2 /2
(
+ VA Q t e( x −σ ) /2
2
)
(6.38)
SG
∞
dx
H4 = ∫ λe
−∞
x 2 /2
( )
+ VA Q t e( x −σ ) /2
2 (6.39)

∞
dx
H5 = ∫λ x 2 /2  ( )
+ VA Q t e( x −σ ) /2
2 (6.40)
−∞ SG e
∞
dx
H6 = ∫λ ( x −σ ) 2
/2
( )
+ VA Q t e( x − 4 σx + 2 σ )/2
2 2 (6.41)
−∞ SG e
Since P vSG = PaSG , we can proceed as before and eliminate the term containing
P vSG from Eq. 6.35:
P aSG (VA Qt )

=
( λSG − λ ) (
{( λ − λ ) H − ( λ − λ ) H }
* *
(6.42)
PI SG 2π − λSG H 6 ) 4 SG 5
λ* λ 1 − F v N O .
Note that in Eqs. 6.34 - 6.42, λ = λN2O and = ( 2
)
Figure 6.9. A-a partial pressure difference for nitrogen when R > 1.
A-a partial pressure difference for nitrogen as a function of σ , the degree of ventilation-
perfusion mismatch, for ΔV = 0, -25, -50, -75, -100 ml/min. These values of ΔV are
associated with a Respiratory Exchange ratio greater than or equal to 1. The red arrow
shows the direction of decrease in ΔV .

Although we have multiplied Eqs. 6.26, 6.27, 6.34 and 6.35 by 563, the
partial pressure of nitrogen in dry air, to produce the curves in Figure 6.8 and
Figure 6.9, the quantitative results are only approximate since our analysis is
based on a second gas present in infinitesimal concentrations. This is equivalent
to the simplifying assumption that VIA FIN2 = VA FAN2 in each gas-exchanging
compartment, an assumption made initially by Rahn and Fenn to enable V Q
lines to be drawn on the oxygen-carbon dioxide diagram (76). Once this no
longer applies, Eq. 6.2 must be modified to include the second gas, giving rise
to 2 equations that must be solved simultaneously for each gas to yield the mean
alveolar and mixed pulmonary end-capillary partial pressures. These
complicated equations must be broken down into a series of expressions for
integration purposes. We have not gone down this path as the other conclusions
drawn from Figure 6.8 and Figure 6.9 regarding the direction of the alveolar-
arterial partial pressure gradient still apply as they are due to the assumption that
P v = P a for nitrogen.
The statement by Canfield and Rahn that “whenever differences in
ventilation-perfusion ratio exist among alveoli of the lung…the mean arterial
inert gas pressure is greater than that of the mean alveolar gas” is borne out in
Figure 6.8 with one important proviso: - the ventilation-perfusion differences
must occur in the presence of a reduction in the system volume i.e. ΔV must be
greater than zero. Thus, the line for ΔV = 0 in Figure 6.8 is associated with no
difference between mean alveolar and mean arterial partial pressure for nitrogen.
Furthermore, if the volume change occurs in the reverse direction, i.e. ΔV < 0,

the direction of the alveolar-arterial nitrogen partial pressure gradient is reversed.
This is illustrated in Figure 6.9.
However, an internal circulation of nitrogen will always occur in the
presence of ventilation-perfusion mismatch whatever the value of ΔV . This
example further demonstrates the importance of taking net gas volume changes
into account when considering gas exchange.
CONCLUSION
In this chapter, we have investigated the effect of net gas volume loss such
as that seen during nitrous oxide anaesthesia, on low concentrations of second
gases present. It was shown that the classical partial pressure-solubility diagram
of West, Wagner and Derks (89) must be modified. At the lower end of the
solubility range, the partial pressure-solubility curve of the second gas in arterial
blood is displaced upwards resulting in a disproportionate increase in the second
gas effect in blood compared with alveolar gas. Uptake of the second gas may
occur against an overall alveolar-arterial partial pressure gradient if the degree
of mismatch and the rate of nitrous oxide uptake are sufficiently large and can
occur with highly insoluble gases such as sulphur hexafluoride but is less likely
with agents such as desflurane and sevoflurane in common use today. We have
also shown that it is virtually impossible for V Q mismatch per se to increase
second gas uptake compared to the equivalent homogeneous lung.
Comparing this with the situation described previously by Canfield and
Rahn (8) for nitrogen in steady-state air breathing, we show that for nitrogen, the
direction of the alveolar-arterial gradient follows the direction of net gas volume
movement.

: Elucidating relevant roles
In previous chapters, we have confirmed the following properties of the
second gas effect:
1. It is greater in blood than in the gas phase,
2. It persists well into the period of nitrous oxide maintenance anaesthesia,
3. It is so powerful that it favours second gas uptake, seemingly against a
partial pressure gradient,
4. It increases as the degree of ventilation-perfusion mismatch increases,
5. It is most pronounced with the low soluble agents in current use.
Yet, how low solubility and ventilation-perfusion mismatch can combine
to improve gas transfer remains unexplained. According to basic concepts of
respiratory physiology, the effect on gas transfer should be the opposite (89, 92).
In this chapter, we use the graphical representation of the log normal distribution
of V Q ratios in the lung developed in Chapter 3 to study more closely the roles
of the first and second gases and how these interact in the presence of V Q
mismatch to produce the SGE. The reader is advised to read Chapter 3 and
Chapter 4 before proceeding. We repeat some of the equations from Chapter 4
as required, in developing the theory.
STEADY-STATE GAS EXCHANGE AS A TWO STEP PROCESS
We have previously shown that steady-state gas exchange offers a
practical model for studying the concentration and second gas effects. In this
Chapter 7: Elucidating relevant roles 111

section we describe a previously unrecognized property of steady-state gas
exchange which can be used to explain how a low solubility and high degree of
V Q mismatch combine to improve gas transfer (47). The property may be
stated as follows: steady-state gas exchange can always be represented as the
sum of two steps: 1. gas exchange at constant volume, followed by 2. gas
exchange on volume correction.
We proceed as follows. Consider an inert gas being exchanged under
steady-state conditions in a single gas-exchanging unit of given VA Q in the
lung. If such a gas is delivered in the inspired gas mixture at a rate VI FI and
removed at a rate VA FA , then the uptake of the gas is VI FI − VA FA . This must
equal the uptake by blood which is given by λQ ( Fc ' − F v ) , where Fc ' and F v
( )
are equal to Pc ' PB − P H 2O and P v (P B )
− P H 2O respectively, with Fc ' the
fractional concentration in the dry fraction of a gas sample in equilibrium with
the capillary blood draining the unit and Pc ' the partial pressure of that gas, so
that:
VIFI − VA FA= λQ ( Fc '− F v ) (7.1)
In the case of the physiological gases O2 and CO2, the expression on the
right-hand side of this equation is different because of the complicated binding
with haemoglobin. Either side of Eq. 7.1 may be taken as the rate of transfer of
gas. Let us therefore denote the rate of transfer of some gas G by the symbol
VG . We may then write:
VG VIFI − VA FA

= (7.2)
112 Chapter 7: Elucidating relevant roles

Consider now this process as taking place in two steps, the first step at
constant volume VI , the second step during the volume correction from VI to
VA . This assumption gives rise to the following pair of equations:
VG (=
01) VI FI − VI F ′ (7.3)
VG (=
12 ) VI F ′ − VA FA (7.4)
In Eq. 7.3, FI and F ′ are the fractional concentrations in dry gas at the
start and end of step 1 and VG ( 01) is the uptake (or output) of the gas in this
step. Similarly, in Eq. 7.4, F ′ and FA are the fractional concentrations in dry
gas at the start and end of step 2 and VG (12 ) is the uptake (or output) of the gas
in this step. It follows from the addition of Eqs. 7.3 and 7.4 by comparison with
Eq. 7.2 that

= VG VG ( 01) + VG (12 ) and so may also be denoted by the symbol
VG ( 02 ) .
Note that the output of Eq. 7.3 becomes the input to Eq. 7.4, i.e. the
output of step 1 becomes the input to step 2. Moreover, this treatment makes no
assumptions about the nature of the relationship governing the interaction
between the gas G and blood and is therefore always applicable to gas exchange
under steady-state conditions. 37 Furthermore, if we add the term d (VFA ) dt to
the right-hand side of Eq. 7.1, where V is the volume of this individual gas-
exchanging unit, we obtain the following equation which now includes the extra
37
Rahn and Fenn used a pathway in explaining the divergence of the gas R-lines on the O2-CO2
diagram (75) but, in doing so, did not apply the principle enunciated here. This is the first time this
principle has been stated explicitly. Note that Kety’s theory of inert gas uptake (see  2.1), only
included step 1 and therefore could not be expected to predict any of the phenomena associated with
step 2.

term to allow for the change in the amount of the gas in the unit as a function of
time:
VIFI − VA F=  ( Fc '− F v ) + d (VFA ) dt

A λQ (7.5)
This manoeuvre allows us to extend the property described in Eqs. 7.3 and
7.4 to the non steady-state, so that it becomes a fundamental property of gas
exchange in individual gas exchanging units. Another way of stating this
property is that it is always possible to find F ′ such that Eqs. 7.3 and 7.4 are
satisfied.
To extend this property to the lung as whole, we must show that it is
permissible to sum the terms on both sides of Eq. 7.5. i.e. we must show that the
following applies:
N N N
j
=j 1 =j 1
∑ (VIFI − VA F=
A) ∑ λ Q ( Fc '− F v ) + ∑ d (VFA )
j
=j 1
j
dt (7.6)
where the subscript j refers to the jth gas-exchanging unit and there are N such
units. The value of each term is to be the value applicable in the jth unit. For
example, VI refers to the inspired ventilation in the jth unit, VI j and Q refers to
the blood flow to the jth unit Q j .
The sum on the left side of Eq. 7.6 requires that the uptake of gas G
from the inspired alveolar ventilation for the whole lung should equal the sum
of the uptakes from the inflow into the individual gas-exchanging units which is
clearly true. The first sum on the right side of Eq. 7.6 requires that the uptake of
gas G by the total pulmonary blood flow should equal the sum of the uptakes
from by the individual blood flows to each unit which again is obviously true.

This leaves the second sum on the right side of Eq. 7.6. For this to be extended
from individual units to the lung as a whole, we must show that:
( ) ( )
N N
d ∑ V j FA j dt = ∑ d V j FA j dt (7.7)
=j 1 =j 1
We can illustrate that this is true by considering the changes in the four-
compartment model of the lung shown in Table 7.1.
Table 7.1. Gas exchange in a lung with 4 compartments
Compartment 1 2 3 4 Lung
Units before 1 5 10 4 20
Units after 2 1 2 2 7
Change +1 -4 -8 -2 -13
The number of units of the gas in each compartment before exchange is shown
in row 2, the number remaining after exchange is shown in row 3, the change in
number of units is shown in yellow in row 4. The sixth column shows what is
happening in the lung as a whole. Units may be number of molecules, moles or
litre at STPD. It can be seen that the sum of the changes for the 4 compartments
in row 4 (in yellow) is equal to the change in the sum for the lung as a whole as
shown in the last entry in column 6 (in green). As the third sum in Eq. 7.6 applies
to the lung as a whole, we conclude that gas exchange in the lung may always
be expressed as a two-component process.

We next attach an appropriate meaning to the numbers 0, 1 and 2 which
G ( 02 )
appear in relation to VG in the equation V= VG ( 01) + VG (12 ) :
0. initial introduction of the gas G under consideration
1. situation after exchange of gas G at constant volume
2. situation after further exchange of gas G on volume correction,
so that (01) denotes step 1, (12) denotes step 2 and (02) denotes the sum of steps
1 and 2. This sequence is now to be used as a template for investigating the
concentration and second gas effects in the system we have used in previous
chapters, i.e. a system consisting of first gas, second gas and insoluble filler gas.
EQUATIONS FOR UPTAKE OF FIRST GAS & DILUENT GAS
The first gas (FG) is assumed to be inert and so will obey Eq. 7.1. For the
diluent gas, which is treated as if it were completely insoluble, we have:
VI (1 − FI ) = VA (1 − FA ) (7.8)
Assuming complete equilibration between the first gas and blood so that
Pa = PA , these equations give rise to a general equation for the alveolar and
arterial partial pressures of the first gas (Eq. 6.3).
EQUATION FOR INSPIRED VENTILATION
It is also possible to obtain the following expression for VI VA :
( VA Q t ) + λ (1 − F v )
(VI VA ) = (7.9)
( VA Q t ) + λ (1 − FI )
With a log normal distribution of V Q ratios, Eq. 7.9 may be expanded to give
the following density function for VI as a function of x = (1 σ ) log ( v q ) :

 
  VA −( x − σ 2)2 2   VA λ ( FI − F v ) 
VI ( x ) =  e  +  
 2π  2π λ (1 − FI ) e( ( ) 2
2
+ VA Q t e( x + σ
x − σ 2) 2
 2)2
   
(7.10)
In Eq. 7.10, the left-hand term in curly brackets generates the distribution of the
expired ventilation while the right-hand term in curly brackets generates the
distribution of the first gas uptake. Both of these distributions are generated as
functions of x = (1 σ ) log ( v q ) . Summed over the whole lung and then
∞
expressed as the equivalent Lebesgue-Stieljes integral, ∫ V ( x) dx
−∞
I yields the
total inspired alveolar ventilation VI . Once λ , VA , Q t , σ and F v are specified,
the integral provides a one-to-one relationship between VI and FI , and hence
allows the calculation of the inspired concentration of the FG associated with the
desired total inspired alveolar ventilation (46). For convenience, we set F v in
Eq. 7.10 equal to 0. Alveolar ventilation VA is set at 4 l/min and pulmonary blood
flow Q t is set at 5 l/min. We may now determine the value of FI required to
 VI − VA and the associated distribution of the

produce any chosen value of ΔV=
FG uptake. Graphs of the FG uptake may now be drawn on the same x-axis as
that used in Figure 3.3.
EQUATION FOR UPTAKE OF SECOND GAS
When present as the sole soluble inert gas, the uptake profile of the second
gas (SG) is also given by the right-hand term of Eq. 7.10 in curly brackets. Since
this gas is present in low concentrations (we use a concentration of 1% in the dry
inspired gas mixture), the FI term in the denominator may be ignored so that
we treat the SG as if it makes no contribution to the volume change. In the

presence of a net gas volume reduction associated with the uptake of FG, the
equation for the uptake of SG must be multiplied by Eq. 7.9, the correction factor
for the difference between the inspired and expired ventilation of each gas-
exchanging unit.
UPTAKE SEQUENCE WITH A FIRST AND SECOND GAS
When second gas uptake occurs in the presence of a soluble first gas, we
set F v for each gas to zero and use the following sequence, which mimics the
steps defined above:
0. Both gases are introduced via the inspired alveolar ventilation to each
gas-exchanging unit
1. The SG is then equilibrated with the blood flow to that unit (step 1)
2. Gas exchange is then completed by allowing the FG to equilibrate with
the blood flow (step 2).
Splitting the gas exchange process into 2 steps in each gas exchanging unit
makes it possible to evaluate separately the effect of the volume contraction
caused by FG uptake in that unit.
DERIVATION OF DISTRIBUTIONS FOR DIAGRAMS
Diagrams will be used to illustrate the processes taking place
simultaneously. It is therefore necessary to derive expressions for the
distributions of inflow, uptake, retention and outflow of first and second gases.
We proceed as follows.

Start by defining the interval on the x-axis to be partitioned into n equal
parts. Since we have specified that P v equals zero for each gas under
consideration, we need only consider gas being brought into the lung by the
inspired ventilation so an acceptable interval is ( μ − 3) ≤ x ≤ ( μ + 3) where μ =
σ 2 . If contributions from mixed venous blood need to be considered, the left-
hand limit must be extended to ( − μ − 3) . If we restrict σ to the interval [0, 2], a
partition on the interval [-4, 4] can be used throughout. Choosing n = 800 will
then provide 101 points in each unit sub-interval, e.g. from -4 to -3, each of the
n intervals will be 0.01 units wide. This will generally provide sufficient
coverage. We now have a partition { xk } of the interval [-4, 4] with n+1 points
such that −4 = x1 < x2 < x3 < ....xk .... < xn < xn +1 = 4 and define the vector
x = [ x1 , x2 , xn , xn +1 ] . For each xk in x and using the notation described in
 7.1, we define the following parameters and the associated vectors which are
listed in Table 7.2

Table 7.2. List of vectors used in obtaining distributions of gas uptake.
Vector Definition Description
x [ x , x , x , x ]
1 2 n n +1
Partition on x-axis with -4 ≤ x ≤ 4
I
V VI1, VI 2 ,VI n −1, VI n  Gas inflow for each x
A
V VA1,VA2,VAn −1,VAn
 
Gas outflow for each x

Q Q1 , Q 2 , Q n −1 , Q n  Blood flow for each x
 IG (01)
V  I FI
V Gas G inflow in step 1 for each x
G
G ( 01)
V VG (01)1 ,VG (01) n  Gas G uptake in step 1 for each x 38
G ( 02 )
V VG (02)1 ,VG (02) n  Gas G uptake in steps 1 & 2 for each x
 A G(01)
V  IG (01) - V
V G (01) Gas G outflow from step 1 for each x
G ( 12 )
V G ( 02 ) - V
V G (01) Gas G uptake in step 2 for each x
38
VG(01) is obtained from Eq. 7.14. Note that the sum of the values for all xi has not been converted
to the equivalent Lebesgue-Stieltjes integral. To do this, we would have to substitute for VIk using Eq.
7.10 in Eq. 7.14. This is far too complicated. Instead, we calculate the volume of N2O taken up for
each value of x using the Lebesgue-Stieltjes form of Eq. 4.8 and add it to the value of VAk calculated
for that value of x from Eq. 7.12. This is far more efficient.

The inspired alveolar gas flow:
VA 2
− ( xk − μ ) 2 VA λ ( FI − F v )
=VIk e +
2π 2π λ (1 − FI ) e( xk − μ ) ( )
+ VA Q t e( xk + μ ) 2  (7.11)
2 2
2
 
 I = VI1,VI2,VIn −1,VIn 
V  
The expired alveolar gas flow:
VA − ( xk − μ )2
VA k =
2
e
2π (7.12)
 A = VA ,VA ,VA
V ,VAn
 1 2 n −1
The pulmonary capillary blood flow:
Q t − ( xk + μ )2
Q k =
2
e
2π (7.13)
 = Q , Q , Q
Q , Q n 
 1 2 n −1
Uptake of gas G in step 1:
λVI k ( FI − F v )
VG (01) k =
(
λ + VI Q t k k ) (7.14)
 G (01) = VG (01)1 VG (01)n +1 
V  
Combined uptake of gas G in steps 1 and 2:
VA λ ( FI − F v )
VG (02) =
{
2π λ (1 − FI ) e( xk − μ ) ( ) }
2
+ VA Q t e( xk + μ )
2
2 2
(7.15)
G ( 02 )
V = VG (02)1 , VG (02) 2 ,VG (02) n −1 , VG (02) n 
The inflow of gas G in step 1 is obtained by multiplying the vector I

V
by the fractional concentration of G in inspired gas i.e.  I G (01) = V

V  I FI G .
The gas retained at the end of step 1 becomes the outflow of the gas G at the end

of step 1: AG (01) = V
V  IG (01) - G (01) . This in turn becomes the
V
inflow of gas G in step 2. The uptake of gas G in step 2 is given by
V G ( 02 ) - V
G ( 12 ) = V G (01) .
Thus, by using vector algebra it is possible to obtain the distribution of any
required parameter. The total uptake of gas over the whole range of x values may
be obtained by integrating the relevant expression using the integral function of
MATLAB but this involves such a complicated expression for the integrand in the
case of Eqs. 7.14 and 7.15 that the much simpler option of applying the
trapezoidal rule (4, 36) has been used instead here and for other calculations,
where appropriate.
As in previous chapters, all programs were written in MATLAB, and where
necessary, equations were solved using the fsolve nonlinear algebraic equations
solver. When required, the integral function was used to integrate improper
integrals. For individual figures involving N2O, values of ΔV were chosen so
as to demonstrate differences when these occur, while also lying within the range
believed to occur clinically during N2O anaesthesia. When the uptake of other
first gases was studied, comparable values of ΔV were used.
DISTRIBUTIONS OF INFLOW, OUTFLOW AND UPTAKE
We begin with graphs showing the inspired and expired ventilations as a
function of x = (1 σ ) log ( v q ) . The inspired alveolar ventilation is obtained
from the sum of the terms in both sets of curly brackets in Eq. 7.10. The expired
ventilation is obtained from the bracketed left-hand term. The resultant inspired
and expired ventilations are shown in the left column of Figure 7.1. The upper

curve in each figure shows how VI varies with x; the lower curve shows how
VA varies with x. The degree of V Q mismatch is increased from A to C. In A,
σ = 0, in B, σ = 1, in C, σ = 2. In each case the curve for VA is symmetrical
about the line x = σ 2 . The curve for VI is symmetrical about the line x = σ 2
only when σ = 0; for higher values of σ it is asymmetrical and bulges to the left.
Figure 7.1. Distribution of alveolar ventilation and individual gas uptake.

LEFT COLUMN: Distribution of inspired alveolar ventilation VI (upper curve) and
expired alveolar ventilation VA (lower curve) as a function of x = (1 σ ) log ( v q ) A: σ
= 0, B: σ = 1, C: σ = 2. In each case, the graph for VA is symmetrical about the line x =
σ 2 (vertical green lines). The yellow-shaded area between the two graphs shows where
gas volume uptake occurs. VI and VA are in l/min. Total expired alveolar ventilation is
equal to the area under the lower curve and is set at 4 l/min. RIGHT COLUMN: Inspired
and expired flows of N2O vs. x. The inspired N2O concentration has been adjusted so as
to produce the same total uptake of N2O in each figure. The yellow-shaded area between
the two graphs shows where N2O uptake occurs and is the same size as in the left column.
The inflow and outflow of 1% sevoflurane is shown in red and is visible at the bottom of
the diagram, just above the x-axis. Note that the y-axis has been shifted to the left in B
and C so that the expired alveolar ventilation is aligned with that in A.
In the right column, we have drawn graphs of the inflow and outflow of
the first gas, N2O ( λ = 0.47) and a second gas, in this case sevoflurane ( λ =
0.65). The graphs for N2O are smaller than those for VI in the left column
because the concentration of N2O has been allowed to vary and has been chosen
to maintain the uptake at 575 ml/min, equal to the value that occurs when σ = 2
and FI N2O = 0.7 , so the volume of N2O taken up is always less than the
simultaneous value of VI which exceeds 4 l/min. As V Q mismatch is
increased, a higher FI N2O is required to maintain the same net volume uptake.
The yellow-shaded area between the inflow and outflow curves is the same
as in the left column both in size and in its location with respect to the x-axis and
is equal to the volume contraction in l/min due to N2O uptake. The red inflow
and outflow graphs for 1% inspired sevoflurane appear at the bottom of this
figure and are so small as to seem insignificant when compared both with the
curves for 70% N2O and the net volume uptake (yellow area). If we are to show
the distribution of SG uptake diagrammatically, it will be necessary to use a scale

on the y-axis that precludes simultaneous inclusion of the FG and the alveolar
ventilation unless these are drawn using an appropriate change in scale. Note
how, as the degree of V Q mismatch increases, the uptake of N2O shifts further
and further to the left, following the movement of the blood flow distribution
(see Figure 3.4). This behaviour is discussed further in  7.13.
SECOND GAS UPTAKE AS A TWO-STEP PROCESS
Let us now consider the uptake of 1% sevoflurane and 70% N2O as a two-
step process in a lung with perfect matching between ventilation and blood flow,
i.e. σ = 0. This is shown in Figure 7.2. In the left panel, sevoflurane exchange is
allowed to occur before the uptake of N2O. The distribution of the inflowing
sevoflurane is shown in blue above the x-axis; that transferred to blood is shown
in red below the x-axis. The distribution of sevoflurane remaining after uptake
is shown in black above the x-axis. The grey area between the black curve and
the x-axis is equal to the volume per minute of sevoflurane which remains after
uptake in step 1. This forms the outflow from step 1 which becomes the inflow
of sevoflurane to step 2. The area of the red region below the x-axis is equal to
the uptake of sevoflurane per minute and matches the area of the white region
between the blue and black curves above the x-axis. The yellow area shows the
distribution of the N2O uptake which is about to occur and the results of which
are shown in step 2.
For reasons of scale discussed above, the yellow area is not equal to the
uptake of N2O per minute but is proportional to it in size. Its value is maintained
constant in the figures that follow. N2O uptake is now permitted to take place
and is accompanied by a contraction in volume proportional in magnitude to the

Figure 7.2. Gas uptake as a two-step process.
Gas uptake shown as a two-step process when ventilation and blood flow are perfectly
matched. Second gas flows are in l/min. Step 1: 1% sevoflurane uptake which takes place
with minimal change in total gas volume. The area between the blue curve and the x-axis
is equal to the sevoflurane inflow. The grey area between the black curve and the x-axis
is equal to the sevoflurane outflow. The red area below the x-axis is equal to the
sevoflurane uptake. The yellow region at the top of the figure is proportional to the
volume contraction which will accompany the N2O uptake in the next step. Step 2: after
N2O uptake has taken place. The associated contraction in volume leads to a further
transfer of sevoflurane to blood (red area below x-axis). The colour of the black curve in
step 1 has been changed to blue as it now represents the sevoflurane input to the second
step. Note how all areas are symmetrical about the y-axis.
size of the yellow area. This produces an increase in the partial pressure of
sevoflurane in the gas phase leading to a further uptake of sevoflurane by blood
shown in red below the x-axis. The sequence is similar to that shown in step 1,
however the black curve in step 1 now becomes the starting point for sevoflurane
uptake in step 2 and is therefore shown in blue. Notice how, in this case of perfect
matching between ventilation and blood flow, the incoming sevoflurane in steps
1 and 2 is perfectly aligned with the uptake of N2O. The right-hand diagram
depicts the SGE which has now been isolated from the contribution to the uptake
of SG at constant volume shown in step 1.

Figure 7.3. Uptake profile of SF6, sevoflurane and acetone.
Uptake profile of sulphur hexafluoride (A), sevoflurane (B) and acetone (C).
Step 1: Distribution of three gases comprising 1% of the inspired gas mixture before
(blue upper curve) and after (lower black curve) equilibration with blood but before
N2O exchange is permitted. The distribution of the absorbed gas is shown in red below
the x-axis and that of the unabsorbed gas is shown in grey above the x-axis. Second gas
flows are in l/min. The yellow shading indicates the distribution of the volume
contraction which is to take place in the next step. This has been scaled to
approximately 1/70 of its actual value so as not to dwarf the volume of each SG.
Step 2: N2O uptake is now allowed to take place. The distribution of SG in the gas
phase at the end of step 1 acts as the starting point for the further uptake (blue curve)
which accompanies the contraction in volume producing the final distribution of SG in
the gas phase (grey area above the x-axis). The extra uptake is shown in red below the x-
axis.

To elucidate further the role of SG solubility and V Q mismatch, the
uptake of several second gases present in the inspired gas mixture in a 1%
concentration will now be examined in the presence of 70% N2O ( λ = 0.47) with
σ = 2. We will consider sulphur hexafluoride, sevoflurane and acetone
( λ = 0.0076, 0.65 and 245, respectively).
Of these gases, only sevoflurane is actually used as an anaesthetic agent –
the other gases are included for the sole purpose of investigating further the role
of solubility in producing the SGE. While there is evidence that highly soluble
gases such as acetone are mainly absorbed in the airways rather than in the
alveoli (see  7.16), we use it here purely as an example of a highly soluble gas
to demonstrate the underlying principles. The process is shown for each gas in
two steps in Figure 7.3 and follows the same pattern as in Figure 7.2, with the
SGE shown in the right-hand column. The largest transfer of SG occurs in Figure
7.3 with sevoflurane. Using these diagrams, we can now investigate the factors
controlling the magnitude of the SGE.
ROLE OF THE SOLUBILITY OF THE SECOND GAS
For the SGE to be maximized, the amount of SG exposed to the contraction
in net gas volume in step 2 must be as great as possible. This gas is shown in
each left-hand diagram in grey. Maximization of the grey area will occur if the
loss of SG to blood during the first step is minimized i.e. if the solubility of SG
in blood, λSG , is as low as possible. Thus, in the case of acetone, so much of the
gas present initially has been transferred to blood during step 1, that there is
almost no acetone left to respond to the volume contraction in step 2. On the
other hand, if the transfer of SG to blood in step 2 is to be maximized, λSG should

be as high as possible. These conflicting requirements for maximization suggest
that for any given set of the other input variables, ( λFG , VA , Qt , σ , FI , F v )
there is a value of λSG for which the SGE is maximized. In the sequence of gases
shown in Figure 7.3, this occurs with sevoflurane. Using the conditions which
apply in Figure 7.3 in Figure 7.4, we have drawn a graph of the total SG uptake
in step 2 vs. log10 ( λSG ) . It can be seen that the maximum occurs for log10 ( λSG )
= -0.41 which corresponds to a value of 0.4 for λSG , consistent with the position
of sevoflurane in the sequence shown in Figure 7.3.
Figure 7.4. Second gas uptake profiles with σ fixed at 2.

Total uptake of SG in step 2 of schematic mechanism shown in Figure 7.2, as a function
of log10 ( λSG ) with an inspired N2O concentration of 0.7 and degree of mismatch given
by σ = 2.
ROLE OF V/Q MISMATCH IN RELATION TO THE SECOND GAS
It now becomes evident that V Q mismatch plays a similar role to that of
the solubility of the second gas, λSG , needing to be as high as possible in step 1

so as to conserve as much of the SG as possible in the first step but as low as
possible in step 2 so as to facilitate maximum transfer of the SG to blood during
the second step. We can therefore expect that for each value of λSG , there is a
value of σ which maximizes the transfer of SG in step 2. Before we can assess
this relationship further, it is necessary to investigate the effect that changes of
σ may have on the disposition of the volume contraction which accompanies
N2O uptake. One fact, however, remains clear - notably that in neither of the two
scenarios discussed so far in relation to V Q mismatch is there any suggestion
that V Q mismatch per se can accelerate rather than hinder gas transfer in the
lung. In step 1, we exploit the fact that a high degree of V Q mismatch impedes
gas uptake while in step 2, we exploit the fact that a low degree of V Q
mismatch improves gas transfer.
EFFECT OF V/Q MISMATCH IN RELATION TO THE FIRST GAS
Not only is the magnitude of the volume contraction important but also its
location in the distribution of V Q values. This is well illustrated in Figure 7.3C
in the case of acetone. Although most of the acetone has been transferred to
blood in step 1, a small amount remains in the high V Q range for exposure to
the volume contraction in step 2. However, the contraction occurs mainly in a
low V Q region (yellow area at the top of Figure 7.3C). It does not overlap the
remaining acetone and there is therefore virtually no SGE at all. In Figure 7.1,
we have already shown that the effect of increasing ventilation-perfusion
mismatch on N2O is to shift the uptake curve to the left. Clearly, the effect of
V Q mismatch on the alignment between the residual SG at the end of step 1

and the contraction about to occur in step 2 must be considered. Factors which
control the position of FG uptake are discussed further in  7.13 below.
Figure 7.5. Second gas uptake profiles as a function of V Q mismatch.
Total uptake of SG in step 2 of schematic mechanism shown in Figure 7.2, as a function
of log10 ( λSG ) with σ varied from 0 to 2 in increments of 0.01. The N2O uptake is kept
constant at 575 ml/min by changing the inspired concentration of N2O. The red curve
represents a value of σ = 0, the blue curve represents a value of σ = 2, as in Figure 7.4.
The direction from red to blue of a vertical line drawn at any given solubility indicates
whether SG uptake in step 2 increases or decreases as σ is increased. Red and blue curves
intersect at points X and Y. To the right of Z, gas exchange occurs almost exclusively in
the airways so no V Q effect will be detectable. The large black arrow points to a small
region where SG uptake increases after first decreasing as σ is increased.
In Figure 7.5 we have graphed the total SG uptake as a function of
log10 ( λSG ) for values of σ ranging from 0 to 2. The volume of N2O taken up
has been kept constant at 575 ml/min by varying the inspired concentration of
N2O. The red curve corresponds to the situation when σ = 0; the blue curve
corresponds to the situation when σ = 2. The two curves cross at the points
labelled X and Y. To the left of X, an increase in V Q mismatch is associated
with an increase in the amount of SG transferred in step 2. Since there is already

a significant degree of SG retention in this solubility range (see Figure 7.3A),
the likeliest cause of the increased SGE is the improved alignment between the
N2O uptake and the SG retained at the end of step 1. To the right of Y, an increase
in V Q mismatch is associated with a similar but smaller increase in the amount
of SG transferred in step 2. Since there can be no improvement in alignment
here, due to the wide separation of the retained gas and the volume contraction
in terms of their position on the x-axis, a separation that only worsens as σ is
increased (see Figure 7.3C), the increased SGE must be due to the effect of V Q
mismatch in increasing the amount of SG retained at the end of step 1. To the
right of Z, gas uptake occurs almost exclusively in the airways and is therefore
not affected by V Q mismatch (see  7.16). Whether malalignment occurs
depends on the distribution of V Q ratios (assumed here to be log normal).
Thus, we conclude that as we move from left to right along the solubility
axis, the importance of correct alignment wanes while that of retention increases.
In the region between X and Y, these processes are not powerful enough to
sustain an increase in the SGE with an increase in V Q mismatch.
A THIRD-ORDER CONTRIBUTION TO SG UPTAKE
In Figure 7.5, the large black arrow points to a small area between X and
Y which has been enlarged in Figure 7.6 in which λSG and σ are shown on the
x-axis and y-axis respectively with the SG uptake on the z-axis. The effect of
increasing the degree of V Q mismatch in this region varies and may involve
an initial decrease in SG uptake followed by an increase when σ is increased
further, as shown by the red arrows. The effect is small and probably
insignificant.

Figure 7.6. 3-dimensional representation of second gas uptake.
3-dimensional representation of region in Figure 7.5 with 0.1 < log10 ( λSG ) < 1.0 showing
how SG uptake first decreases as σ is increased from 0 to 1 then increases as σ is

increased further from 1 to 2 (red arrows).
LOCATION OF UPTAKE IN THE V/Q SPECTRUM
The location of the FG uptake and how it is affected by increases in V Q
mismatch is clearly of importance in determining the degree to which the SGE
finds expression. To understand this better, we first examine the effect of an
increase in σ on the uptake distribution of a gas present in very low
concentrations in the inspired gas mixture. We then examine what will happen
as we increase the inspired concentration. The effect of increasing σ is
demonstrated for a low concentration of a highly insoluble gas in Figure 7.7A
using sulphur hexafluoride. When σ = 0, there is no mismatch and the uptake is
distributed symmetrically about the y-axis (dark blue curve).
As the degree of V Q mismatch is increased, the uptake profile shifts
further and further into the low V Q zone. We can describe sulphur
hexafluoride uptake as perfusion-dependent when V Q mismatch increases.
The situation for a highly soluble gas is illustrated in Figure 7.7B using
methoxyflurane. It can be seen that the uptake profile shifts further and further

Figure 7.7. Ventilation and perfusion dependent uptake of inert gases.

Fractional uptake of 1% inspired concentration of gas as a function of
x = (1 σ ) log ( v q ) for increasing values of σ . A: sulphur hexafluoride ( λ = 0.0076),
B: methoxyflurane ( λ = 13) and C: hypothetical gas with λ = VA Q t . σ = 0 (blue), σ =
0.5 (red), σ = 1 (green), σ = 1.5 (purple), σ = 2 (magenta). The graphs in A and B
become asymmetrical as σ is increased so the vertical line in each graph is located at the
median value of y. The uptake profile of sulphur hexafluoride shifts to the left as σ
increases; an example of perfusion-dependent uptake; the uptake profile of
methoxyflurane shifts to the right as σ increases; an example of ventilation-dependent
uptake (see direction of arrows in A and B).
into the high V Q zone so is ventilation-dependent. These shifts are indicated
by the direction of the arrow in each of these figures. It is possible for the uptake
of a gas to be neither perfusion-dependent nor ventilation-dependent as σ is
increased. This is shown in Figure 7.7C and occurs when λ = VA Q t which in
our case is equal to 0.8. Note that we have used the fraction of the total uptake
as the variable on the y-axis rather than its absolute value as we are interested
here in the location of the uptake rather than the fact that the uptake must
decrease as V Q mismatch increases. Moreover, we know from  3.11 that
trends in a distribution plotted against (1 σ ) log ( v q ) reflect similar trends in
that distribution when plotted against log (V Q ) .
From Eq. 7.10 it is possible to predict the effect of increasing the
concentration of any gas in the inspired gas mixture. As FI increases, the
solubility of the gas effectively becomes λ (1 − FI ) so, for example 70% N2O
behaves like a gas with λ = 0.47*(1 - 0.7) = 0.14, present in very low
concentrations. As a result, the uptake distribution curve always shifts to the left
for any gas as its inspired concentration is increased. N2O-driven SGEs therefore

involve compartments with a low V Q where a low ventilation results in more
powerful concentrating effects on alveolar gas and arterial blood.
The concept of effective solubility may be extended to include the
physiological gases. However the solubility of these gases is not constant but
varies according to numerous factors which must be taken into account by using
a suitable algorithm for their interaction with haemoglobin, for example (14).
For gases present in very low concentrations, the solubility and effective
solubility are essentially the same.
EFFECT OF VENTILATION-DEPENDENT FG ON SG UPTAKE
It follows from the above discussion that the magnitude of the SGE might
be expected to differ if a FG is used whose uptake is ventilation-dependent as
the degree of V Q mismatch increases. Although it is no longer used clinically,
the effect of diethyl ether on sevoflurane in step 2 is compared with that of N2O
in Figure 7.8. 39 It can be seen that the SG uptake is reduced in Figure 7.8B
(diethyl ether) compared with Figure 7.8A (N2O). The shape of the inspired SG
distribution (blue curve) is slightly different in Figure 7.8B than in Figure 7.8A
because it is determined by the distribution of VI which is affected by the
different locations of the FG uptake in the case of ether and N2O (yellow region
at top of each diagram).
39
Note that according to Anderson et al. (1, 2), diethyl ether absorption occurs to a large extent in the
airways, rather than in the alveoli but we use it here to illustrate the possible effect of a ventilation-
dependent FG by assuming that it is exchanged solely at alveolar level (see  7.16). This assumption
has also been made with respect to methoxyflurane in Figure 7.7B.

Figure 7.8. Effect of volume contraction with N2O and diethyl ether.
Effect of producing a volume contraction of 575 ml for sevoflurane in step 2 when σ =
2; A: 70% N2O, B: 25% diethyl ether. Flows are in l/min.

THE CONCENTRATION EFFECT
We now apply these principles to the first gas, whose uptake is itself
accelerated by the contraction in volume which accompanies its uptake. This is
known as the concentration effect (CE) and was first described by Eger (18, 21).
Treating the first gas uptake as a two-step process as before, with the first
step at constant volume, while the second step incorporates all volume changes,
we obtain the results shown in Figure 7.9 for N2O ( λ = 0.47) and diethyl ether
( λ = 12).
Figure 7.9. Two-step analysis of the concentration effect.

Two-step analysis of the concentration effect for N2O (A) and diethyl ether (B). An uptake
volume of 575 ml/min is achieved with an inspired concentration of 70% N2O and 25%
diethyl ether. Gas flows are in l/min. The area of the red region in step 2 is equal to the
extra volume per minute of each gas taken up because of the contraction in volume.
As ether is far more soluble in blood than N2O, it requires an inspired
concentration of approximately 25% to produce the same net volume change as

70% N2O when σ = 2. The left diagram in each pair shows the result of step 1 -
uptake at constant volume. The gas transferred to blood is shown in red below
the x-axis. Its removal must be accompanied by a decrease in partial pressure in
the gas phase, which is corrected in step 2, during which an additional amount
of gas is transferred to blood.
The sum of the red areas in the left and right diagrams in Figure 7.9A and
Figure 7.9B is identical but the volumes arising in each step differ. According to
our previous argument in relation to the SGE, it is the transfer in step 2 that is
attributable to the concentration effect. The effect of V Q mismatch on the
alignment of the contraction with the gas retained at the end of step 1 is no longer
a factor but its effect on the retention of FG in step 1 occurs as before. It can be
seen that for equal net gas volume uptake, the CE is greater for the less soluble
agent N2O.
Figure 7.10 shows similar plots to Figure 7.5 and Figure 7.6. It clearly
indicates that the CE always increases as the degree of V Q mismatch increases
with the increase being greater at lower solubilities. Thus looking at the gas with
log10 ( λFG ) = 1, (i.e. λ = 10) in Figure 7.10B, the CE increases as we move left
to right from 0 to 2 along the V Q mismatch axis. This increase is accentuated
more and more as we move left along the solubility axis and is maximal at the
left end of the solubility axis when log10 ( λFG ) = -0.8, (i.e. λ = 0.16). There are
no decreases in CE similar to those between X and Y with the SGE in Figure 7.5
because malalignment no longer plays a part in gas exchange.

Figure 7.10. Influence of solubility on step 2 of concentration effect.
Uptake of FG in step 2 of schematic sequence. A: Uptake shown in Figure 7.5 as a
function of log10 ( λSG ) with σ varied from 0 to 2 in increments of 0.01. The red curve
represents a value of σ = 0, the green curve represents a value of σ = 2. The upward

direction from red to green of a vertical line drawn at any given solubility suggests that
the FG uptake in step 2 always increases as σ is increased. The red and green curves do
not intersect because malalignment is no longer an issue. B: 3-dimensional representation
of Figure 7.10A with σ on the x-axis, log10 ( λSG ) on the y-axis, and uptake of the FG in
step 2 on the z-axis. This confirms that the concentration effect for any gas always
increases as the degree of V Q mismatch increases.

ABSORPTION OF GAS IN AIRWAYS RATHER THAN ALVEOLI
It has been shown by Anderson et al. that with λ > 10, gas exchange begins
to take place in the airways, as well as in the alveoli (1, 2). V Q scatter therefore
does not play the same role as it does with gases of lower solubility. The effect
of V Q mismatch in increasing the SGE shown in Figure 7.5 for gases with λ
greater than 10 will therefore only be of importance for the alveolar component
of gas exchange and may not be significant clinically. This serves to further
explain why the effects of V Q mismatch on the SGE were more likely to have
been demonstrated experimentally with the anaesthetic liquids of low solubility
currently in use.
The transition from alveolar to airway gas exchange progresses as λ
increases. With λ greater than 100, exchange takes place almost exclusively in
the airways so that our use of acetone in the discussions above is purely
hypothetical. Because of the wide separation between the location of the gas that
remains at the end of step 1 and the location of the contraction of volume in step
2, acetone illustrates the effect of malalignment more clearly than is possible
with a gas such as halothane or methoxyflurane. This is the rationale for using it
in our examples above.

AN ALTERNATIVE DEFINITION OF THE AUGMENTATION RATIO
We have previously expressed the magnitude of the SGE in terms of an
augmentation ratio (AR) defined by Epstein et al. (24) as the partial pressure of
the SG in the presence of the FG, divided by that which would have existed in
the absence of the FG, with all other variables kept constant (45, 46). This
definition is useful because it can be applied in experimental situations.
However, there is no obvious way of expanding the definition to express the
magnitude of the concentration effect. A more useful definition distinguishes
between the combined uptake of gas in steps 1 and 2, divided by the uptake in
step 1, i.e. the portion of uptake that is independent of the volume contraction.
This second definition is easily extended to the concentration effect. In the case
of the SGE, it ensures that the divisor is based on the number of molecules of
gas in VI FI rather than the number in VA FI , and therefore prevents the SGE
from being overestimated.

CONCLUSION
We have demonstrated that steady-state gas exchange in the lung is always
equal to the sum of two components: exchange at constant volume followed by
exchange on volume correction. This applies irrespective of the way the gas
dissolves in blood i.e. it applies whether we are dealing with the physiological
gases which combine with haemoglobin or those gases that simply obey Henry’s
Law. This property of steady-state gas exchange has not been explicitly stated in
the past. Using this property, we have investigated the transfer of gases in the
lung using a two-step model which allows the contribution from net gas volume
contraction, which occurs in step 2, to be separated from gas transfer at constant
volume, which occurs in step 1. Gas uptake in step 2 is identifiable as due to the
concentration or second gas effect depending on whether the gas being studied
contributes significantly to the contraction in volume. The magnitude of these
effects depends on sufficient inflow of the gas to step 2, which itself depends
critically on how much gas is retained after uptake in step 1. Both the effective
λ′ λ (1 − FI ) and the degree of V Q mismatch, σ affect uptake in

solubility, =
each step according to the following table:
Table 7.3. Factors affecting gas uptake in each step of a mechanistic model of gas exchange.
Factor Step 1 Step 2
Low effective solubility Favors retention in gas phase Impedes gas transfer to blood
High effective solubility Favors loss of gas to blood Favors gas transfer to blood
Low degree V Q mismatch Favors loss of gas to blood Favors gas transfer to blood
High degree V Q mismatch Favors retention in gas phase Impedes gas transfer to blood

When dealing with the SGE, an additional factor is the proximity of the
effective solubilities of the first and second gases. This controls the extent to
which the SG retained at the end of step 1, is exposed to the contraction in gas
volume in step 2 and is very much dependent on the distribution of V Q ratios.
The information in Table 7.2 allows us to conclude that both solubility and V Q
mismatch act to increase uptake in step 2 by decreasing uptake in step 1.
While it is useful to distinguish the roles of alignment and retention in
producing the SGE, the contribution of V Q mismatch to increased retention
for a second gas whose solubility in blood exceeds 10 may not be as important
because airway exchange for such gases becomes progressively more important
than alveolar exchange as λ increases. This means that the effect of V Q
mismatch on the SGE to the right of point Y in Figure 7.5 may not be
demonstrable clinically.
We have introduced the terms perfusion-dependent uptake and ventilation-
dependent uptake to describe the behaviour of inert gases as V Q mismatch
increases, according to whether λ is less than or greater than VA Q t . This
represents another new property of inert gases albeit one that may change with
the values of alveolar ventilation and pulmonary blood flow. It indicates whether
the uptake of an inert gas, present in low concentrations in the inspired gas
mixture, moves towards an area of lower or higher V Q as the degree of V Q
mismatch increases in the lung. However, as the inspired concentration of any
inert gas is increased from these low levels, its uptake always moves to a lower
V Q region as indicated by the formula for effective solubility given above, i.e.
it becomes more perfusion-dependent.

: Influence of Respiratory Pattern
INTRODUCTION
The astute reader will have noticed by now that apart from a brief mention
on page 68, the terms constant inflow and constant outflow have not been
mentioned in this thesis since Chapter 2. It should be recalled that in constant
outflow, it is the total expired ventilation which has its value pre-determined at
some fixed rate (we set it at 4 l/min) while in constant inflow, it is the total
inspired ventilation which has its value pre-determined at the same fixed rate
(see Figure 2.8). With these constraints applied, the inflow of fresh gas is always
less in the constant inflow model at any stage of the washin process. Hence it
comes as no surprise to find that with the constant inflow model, the graph of
alveolar/inspired concentration as a function of time lags behind that for the
equivalent constant outflow case for any gas undergoing transfer from lungs to
blood (58). So far in this thesis, we have been using the constant outflow
equations because of their greater simplicity. In this chapter, we examine the
discrepancy between the two extreme patterns, in the first instance by using the
methods of Chapter 5, and then by using the technique developed in Chapter 7.
AUGMENTATION RATIOS FOR CONSTANT INFLOW
As with the constant outflow case, it is possible to produce equations for
the augmentation ratio in the constant inflow case. These are cumbersome and
complicated. I do not propose to reproduce them here, as a full derivation of the
equations for concentration and uptake is given in the next section. However,
Chapter 8: Influence of Respiratory Pattern 145

Figure 8.1. Persistent second gas effects with constant inflow.
Persistent second gas effects with constant inflow for desflurane, isoflurane and diethyl
ether as a function of the proportion of washin of nitrous oxide for different values of σ
. The SGE is expressed as an augmentation ratio (AR), defined as the partial pressure of
the second gas in the presence of nitrous oxide, divided by that which would have existed
in the absence of nitrous oxide, with all other factors kept constant. Upper three diagrams,
SGE in the gas phase; lower three diagrams, SGE in blood. In each case, the augmentation
ratio is plotted on the y-axis as a function of the washin ratio. The long arrow in each
diagram indicates the direction of increasing σ values from 0 to 2 in increments of 0.25
units. The line for σ = 0 is the same for each second gas.
graphs equivalent to those in Figure 5.1 are shown above. These again illustrate the
persistent nature of the second gas effect, even as uptake of the first gas, nitrous oxide,
falls to maintenance levels. They also show the effect of ventilation-perfusion
146 Chapter 8: Influence of Respiratory Pattern

mismatch and second gas solubility previously described for the constant outflow case.
The main features may be summarized as follows:
1. Divergence of the lines from the point (1,1), the point on the x-axis
representing complete washin of the nitrous oxide, at which time expired
alveolar ventilation becomes equal to inspired alveolar ventilation and
the SGE disappears completely.
2. The line for σ = 0 , which represents a homogeneous lung, is the same
for both gas phase and blood phase, as expected for a “single
compartment” lung with full equilibration between gas and blood phases
but is no longer independent of the solubility of the second gas in blood.
3. The effect of increasing σ and hence the degree of V Q mismatch is
seen to be opposite for the blood and gas phases. In blood, an increase in
σ increases the augmentation ratio, while in the gas phase, the
augmentation ratio decreases. In blood, this amplification is greatest for
the least soluble gas, desflurane, and least for the most soluble gas,
diethyl ether, while in the gas phase, the opposite is true.
The effects are not as great as in Figure 5.1. This partly because of the
restriction placed on gas inflow but also because, as was alluded to in  7.17,
augmentation is overestimated in the case of constant outflow because the
second gas molecules brought in with the extra inflow of gas are not included in
the reference. This involves a correction of VA VI which must be applied to the
results in Figure 5.1. This correction did not become apparent until the idea of a
two-step model of gas uptake was conceived almost 18 months later. The

implications of the two-step model when applied to constant inflow are the next
subject to be investigated in this chapter.
THE TWO-STEP MODEL APPLIED TO CONSTANT INFLOW
A more useful comparison of the two extreme respiratory patterns may be
obtained by comparing the SGE when all factors except the respiratory pattern
are kept the same. To achieve this, we utilize the two-step model developed in
Chapter 7. The following derivation should indicate to the reader how one deals
with constant inflow mathematically.
8.3.1 EQUATIONS
First Gas
We begin, as in  4.3 with the following two equations which describe the
steady-state uptake of the first gas in a single compartment model of the lung:
VIFI − VAF= λQ t ( F − F v ) (8.1)
(1 − FI )VI =
(1− F ) VA (8.2)
From Eq. 8.2 we have the following expression for VA :
VA =
(1 − FI ) VI (8.3)
(1 − F )
Letting ψ = VI ( λQ t ) , we obtain the following quadratic equation 40 for the
alveolar concentration F , of the first gas at equilibrium:
40
Eger also solved a quadratic expression for the constant inflow case, describing it as cumbersome.
With minor exceptions, he virtually ignored this case subsequently whenever he discussed the
concentration and second gas effects (19, 22). Mapleson incorporated both cases into his electric
analogue and gave equal weight to each in subsequent discussions (57).

F 2 − ( ψ + 1 + F v ) F + ψFI + F v =0 (8.4)
with the general solution:
(1 + ψ + F v ) ± (1 + ψ + F v ) − 4 ( ψFI + F v )
2
F= (8.5)
2
Only the negative root applies, otherwise F may exceed 1. If F v = 0, we have
the following simpler solution:
(1 + ψ ) − (1 + ψ )
2
− 4ψFI
F= (8.6)
2
The uptake of the first gas is obtained from either side of Eq. 8.1, and with F v
= 0 is given by:
VFG
=
λQ
2 {
(1 + ψ ) − (1 + ψ )
2
− 4ψFI } (8.7)
In a lung model consisting of n parallel compartments, each with its own inspired
ventilation Vj and blood flow Q j , where j = 1 to n, and defining ψ j = Vj ( λQ )
j
we have the following expression for the total uptake of the first gas:
 λ n  
VFG   ∑ Q j (1 + ψ j ) − (1 + ψ )
2
= j − 4ψ j FI  (8.8)
 2  j =1  
Converting to Lebesgue-Stieltjes integrals, and given that V=

A VI − VFG we have
the following expression for VA , equivalent to Eq. 6.6:
V0  λQ 0 
=VA II −   IU (8.9)
2π  2 2π 

2
∞  σ
− x −  2
II = ∫e
−∞
 2
dx (8.10)
{(1 + g ( x)) − }
2
∞  σ
− x +  2
(1 + g ( x) )
2
=IU
−∞
∫e  2
− 4 FI g ( x) dx (8.11)
Here I I is a Lebesgue-Stieltjes integral associated with total gas inflow, I U is a
Lebesgue-Stieltjes associated with first gas uptake. Letting V0 be the total
inspired alveolar ventilation (taken as 4 l/min), Q 0 the pulmonary capillary blood
flow (taken as 5 l/min), η = V0 λQ 0 and letting μ = σ 2 , g ( x) is given by:
2
e −( x − μ ) 2
g ( x) η=
= 2 ηe 2 μx (8.12)
−( x + μ ) 2
e
so that I U becomes:
∞
 
(1 + ηe ) − (1 + ηe ) 2 μx 2
2
−( x + μ ) 2
IU ∫e
2 μx
= − 4ηFI e 2 μx  dx (8.13)
−∞  
Before attempting to integrate the expression for I U , it is necessary to
multiply the first exponential term through the curly-bracketed expression,
otherwise attempting to integrate the term e 2 μx will produce an error as
x → ±∞ . This produces the following expression for I U :
IU = I 1 - I2 (8.14)
∫ (e ) dx
∞
2 2
−( x + μ ) 2
+ ηe −( x − μ ) 2
=I1 (8.15)
−∞
(e ) − 4ηF e (
∞
2
−( x + μ ) 2
2
−( x − μ ) 2
2
− x2 + μ2 )
I2 = ∫
−∞
+ ηe I dx (8.16)

Using our final version of Eq. 8.9, it is now possible to determine the
inspired concentration of the first gas, FI , which corresponds to a specified first
gas uptake for any value of σ , in much the same way as was done in  7.3.
Moreover, Eq. 8.9 provides the distribution of the gas inflow and outflow from
the lung and gas uptake by blood in a similar manner to that used to determine
the distribution of the gas inflow in  7.6:
V − x − μ 2
VI ( x) = 0 e ( )
2
(8.17)
2π
 λQ 0
(
  − ( x + μ )2 2
) ( ) ( )
2
2 2 2 − x2 + μ2
VFG ( x) 
−( x − μ ) 2 −( x + μ ) 2 −( x − μ ) 2
=   e + ηe − e + ηe − 4ηFI e 
 2 2π  
(8.18)
V=
A( x ) VI ( x ) − VFG ( x ) (8.19)
Second Gas
For the second gas, which is present in very low concentrations so that
FI SG  1 , we have the following equation which is similar to Eq. 8.1:
VIFI SG − VAF=
SG λSG Q t ( FSG − F vSG ) (8.20)
This gives the following solution for the uptake of the second gas:
λSG
=VSG
λSG + VA Q
(VI FI SG − VA F vSG ) (8.21)
When F vSG = 0, we have, as before the following uptake as a function of x in
step 1:
λSG FI SG
VSG (01) = VI ( x) (8.22)
λSG + VI ( x) Q ( x)

And the following uptake as a function of x in steps 1 and 2:
λSG FI SG
VSG (02) = VI ( x) (8.23)
λSG + VA( x) Q ( x)
We now construct the vectors shown in Table 7.1, using the new definitions for
VI ( x ) , VFG and VA ( x ) . These are used to generate results similar to those of
Figure 7.3.
The result is shown for 1% desflurane in Figure 8.2 in which we have
combined both steps into one diagram in which we compare constant outflow
(upper diagram) and constant inflow (lower diagram). The desflurane inflow into
step 1 is shown in stippled green. The region between the stippled green line and
the solid blue line corresponds to the uptake of desflurane in step 1. The region
between the solid blue line and the solid black line corresponds to the uptake of
desflurane in step 2; this distribution is also shown in red below the x-axis. The
left vertical black line is positioned at the maximum in the N2O uptake; the right
hand vertical black line is positioned at the maximum in the distribution of the
desflurane remaining at the end of step 1 (grey coloured region).
It can be shown that when all other factors are kept equal, the main effect
of the constant inflow respiratory pattern is to restrict the migration of the nitrous
oxide uptake to the left when it attempts to follow the movement of blood flow
in that direction as σ is increased (see  7.7). This restriction results in a greater
degree of overlap between the second gas retained at the end of step 1 and the
contraction in volume that takes place in step 2. Although not shown here,
expressed as an Augmentation Ratio, the effect can be shown to be greatest for
the least soluble agent, sulphur hexafluoride and virtually non-existent, as in
previous simulations, for acetone.

Figure 8.2. Comparison of constant outflow and constant inflow.
Comparison of constant outflow and constant inflow respiratory patterns for 1%

desflurane with σ = 2. Upper panel: constant outflow, lower panel: constant inflow. The
solid black vertical lines are drawn at the x-value associated with the maximum uptake of
N2O and that of the maximum in the distribution of the desflurane retained at the end of
step 1 (grey region). The effect of malalignment manifests itself because of the
assumption here of a log normal distribution of ventilation and perfusion.

Plotting the Augmentation Ratio as defined in  7.17 for constant inflow
and constant outflow for second gases with solubilities in the range 0.01-100, we
obtain the graphs shown in Figure 8.3. These confirm the predictions based on
Figure 7.3 and Figure 8.2, that augmentation increases for both constant outflow
and constant inflow as the solubility of the second gas decreases, with the effect
more marked for constant inflow, the difference increasing as the degree of V Q
Figure 8.3. Augmentation ratios for constant inflow and constant outflow.
Augmentation ratios for constant inflow and constant outflow with σ = 1 and σ = 2. λSG
has been varied from 0.01 to 100. The variable on the x-axis is log10 ( λSG ) . Inspired
ventilation has been made the same for constant inflow and constant outflow.

mismatch increases with σ going from 1 to 2. As σ → 0 , the curves for constant
inflow and constant outflow come closer together and eventually coincide.
Moreover, given the comments in  7.16, the V Q effect is likely to become
progressively less evident as λ increases beyond 10.
CONCLUSION
Thus, we conclude that when all other factors except V Q mismatch are
eliminated, the constant inflow pattern is associated with a greater SGE than the
constant outflow pattern provided the solubility of the second gas is low enough
and the degree of V Q mismatch sufficiently high to unmask this trend.
Otherwise, the two patterns produce identical augmentation. The cause of this
difference is the better alignment between the SG retained at the end of step1
and the contraction in volume taking place in step 2. The improvement in overlap
arises because the distribution of nitrous oxide uptake is restricted as it tries to
follow the movement of the perfusion distribution to the left as the degree of
V Q mismatch is increased.

: Conclusions
I set out initially to try and prove Philip Peyton was wrong in his
assertion that ventilation-perfusion mismatch can actually improve gas uptake.
The very idea seemed preposterous! Yes, there was experimental evidence
supporting the notion but I believed that other factors might have been at play
and that these would become apparent in time. The theoretical basis for his
explanation of the observations during clinical anaesthesia also seemed suspect
to me. Like Eger (18, 21), Peyton had used a model with many built-in features
(68-70) – dead space, shunt, absorption atelectasis, hypoxic vasoconstriction,
interaction of O2 and CO2 with haemoglobin and the two extreme patterns of
ventilation first identified by Frumin et al. (31) (termed constant inflow and
constant outflow by Alex Robertson). From my own experience working with
Robertson, I could appreciate the importance of peeling off every unnecessary
layer to simplify a problem as much as possible. The more complicated the
model, the greater the likelihood of a mistake buried deep in computer code.
The search for a suitable model ended when I came across an old
manuscript by Colburn, Evans and West (10). Their work had lain dormant for
over 40 years. Although it focused on a two-gas system, the other gases merely
behaving as if they were an insoluble filler acting purely as a vehicle for the gas
under study which in anaesthetic terminology becomes the first gas, it was
immediately apparent to me that a second gas could be added to the system, in
δx amounts, without interfering with the basic equations of gas exchange. This
relationship, a large concentration of a first gas in the inspired gas mixture
together with a second gas present in very low concentrations, mimics the
Chapter 9: Conclusions 157

clinical situation when nitrous oxide anaesthesia is supplemented with a volatile
agent. The manuscript proved to be a veritable goldmine as it contained a
derivation of a log normal distribution of V Q ratios which could be adapted to
provide an efficient way of varying the degree of V Q mismatch, one of the key
variables involved in the study.
It soon became obvious that Peyton’s conclusions were sound, prompting
further investigations of this hitherto poorly-understood phenomenon. The
results are described below as the conclusions of each chapter of this thesis are
presented.
Chapter 1 acts as an introduction explaining my interest in this subject.
In Chapter 2, previous work on the concentration effect and second gas effect
is described and the published work suggesting the phenomenon of
acceleration in the presence of increased degrees of V Q mismatch, of second
gas uptake when modern volatile agents are used together with nitrous oxide, is
presented. The important lesson from this chapter is the following explanation
of the cause of concentration and second gas effects:
Uptake of significant volumes of gas concentrates each remaining gas in a smaller volume;
this increases its partial pressure and accelerates its uptake. In the case of CO2, its elimination
from blood is slowed.
The preoccupation with the flawed diagram produced by Eger to explain his
mathematical model has hopelessly confused discussion of this subject by one
author after another. Eger’s reluctance to abandon the diagram, even after Bill
Mapleson and I showed it to be incorrect, continues to contribute to the
uncertainty surrounding the subject.
158 Chapter 9: Conclusions

Chapter 3 introduces the form of log normal distribution to be used in this thesis.
This is an alternative to the method described by John West. In some ways, it is
superior to the standard technique in that it can be applied as a continuous
distribution, rather than imposing a compartmental approach to gas exchange.
The use of (1 σ ) log ( v q ) as the variable on the x-axis and v or q on the y-axis
produces a graph whose shape and size remains unchanged as σ is varied,
providing a stable platform for studies of the concentration and second gas
effects. The transformation of sums to Lebesgue-Stieltjes integrals is a natural
progression which, when applied, allows one to take full advantage of the
powerful functions available in MATLAB, the computer program recommended
by my colleague Ranjan Dash, that was used in most of the simulations. These
innovations may prove useful to other workers in the future.
Chapter 4 introduces the basic mathematical equations of gas exchange to be
used throughout. The reason for preferring steady-state equations is clarified.
Although the equations themselves are well known, the technique of tagging a
second gas onto these equations has not been suggested before. The minimum
number of components necessary for a study of the SGE are identified. These
are: a first gas which produces the volume changes whose effects are being
studied, a second gas present in minute concentrations and an insoluble filler
gas that acts as a vehicle for delivering the first and second gases.
Chapter 5 addresses the issue of persistent second gas effects. It is shown that
not only does the SGE persist even as N2O uptake falls to maintenance levels
but that as V Q mismatch increases, the partial pressure of the currently used
volatile agents actually increases in blood but decreases in the gas phase. This
has an obvious clinical corollary, namely that MAC readings, based on the

analysis of end-tidal gas, must underestimate the depth of anaesthesia. These
results are all consistent with Peyton’s claims.
Chapter 6 reassesses theoretical work emanating from the San Diego group,
taking into account previously ignored effects of volume change. The main
finding is the possibility of partial pressure reversal during gas uptake. By this
we mean that second gas uptake seems to occur against a partial pressure
gradient between alveolar gas and arterial blood. This is predicted for agents of
very low solubility with a large contraction in gas volume, in the presence of a
significant degree of V Q mismatch. The situation with nitrogen when
breathing air is revisited and shown to be more complicated than described
originally by Canfield and Rahn.
Chapter 7 brings together the techniques used in previous chapters to find a
plausible explanation for the observed phenomena. We derive a hitherto
unreported property of gas exchange: gas exchange in the lung can always be
represented as the sum of two components - gas exchange at constant volume
followed by gas exchange on volume correction. This property permits us to
split gas uptake into two steps. Uptake of the SG is permitted to take place in
both steps but FG exchange is only permitted in step 2 where it produces the
volume correction. In this way, the contribution to SG uptake due to volume
contraction is isolated from other contributions. It becomes obvious that the
amount of SG retained at the end of step 1 determines the SG exposure to the
contraction in step 2 and that this is the primary site where SG solubility and
V Q mismatch exert their effect on the SGE. In addition, the degree of V Q
mismatch can exert an influence through its impact on the alignment of the
contraction in volume with the remaining SG. Thus, retention and alignment

are identifiable as the important factors controlling the contribution of volume
contraction to gas uptake and both agent solubility and V Q mismatch exert
their effects through these two factors. In the case of the first gas, retention
affects the magnitude of the concentration effect but malalignment plays no
role. As part of our investigation, we have characterised the uptake of gases as
perfusion-dependent or ventilation-dependent as V Q mismatch increases,
another property of gas uptake not previously identified.
Chapter 8 returns to the two extreme respiratory patterns. The greater
complexity of the constant inflow case is demonstrated mathematically. The
restriction of gas inflow in this case means that there are always fewer
molecules of second gas available to respond to a contraction in gas volume.
With respect to the SGE, when this advantage of constant outflow is
eliminated, constant inflow is shown to be more effective in responding to
contraction because it produces better alignment but only for agents with a low
solubility in the presence of high degrees of V Q mismatch.
The work in this thesis is a direct continuation of that which came out of
San Diego from West’s group in the 1970s. Until the anomalous behaviour of
the SGE was predicted by Peyton et al. in 2001 (68-70), and the existence of
persistent SG effects demonstrated thereafter (34, 71), there was no real
incentive for Respiratory Physiologists to continue their research in this area.
The golden epoch of physiology in which the secrets of V Q mismatch were
being unlocked, was over. It began in the laboratory of Wallace Fenn, Hermann
Rahn and Arthur Otis in Rochester, New York during the World War II, and was

continued after the war by this group and that of Richard Riley of The John
Hopkins School of Medicine and André Cournand at Bellevue Hospital in New
York City. Leon Farhi joined the Rochester group, which moved to Buffalo, New
York in 1956, where he was later joined by Albert Olszowka. Meanwhile, John
West established the School of Medicine at The University of California, San
Diego in 1969 and led a research team which included Peter Wagner and John
Evans. The epoch came to a close with the successful development of the
Multiple Inert Gas Elimination Technique by West’s group in the 1970s and
early 1980s. When Peyton’s predictions were published in the Journal of Applied
Physiology, most of the members of these teams had either died or redirected
their attention to other more pressing areas of Pulmonary Physiology. Thus, it
was left to the Anaesthetic fraternity to sort out the mechanism that underpins
the seemingly impossible effect of V Q mismatch in enhancing gas uptake.
The place of this research remains unclear. With gaseous anaesthesia
facing an uncertain future, it may prove to be of passing interest. Perhaps we
should be grateful that Peyton discovered the effect of V Q mismatch on gas
uptake when he did. Otherwise, what followed may never have come to light.
The simulations in this thesis could be extended from the steady-state to
more realistic models of anaesthetic uptake. Gas Man®, a currently available
computer simulation tool does not facilitate a study of the effects described in
this thesis. Philip Peyton has repeatedly urged me to undertake the production of
a more realistic simulation tool. Knowing that the underlying model will require
the simultaneous solution of first order differential equations for oxygen, carbon
dioxide, nitrogen, nitrous oxide and a volatile anaesthetic agent as well as a
balance equation for gas volume in up to 100 parallel compartments represents

a daunting task 41, I have always politely declined. Moreover, the change in FRC
that inevitably accompanies induction of anaesthesia has not, to the best of my
knowledge, been satisfactorily described in a computer algorithm.
Notwithstanding this, I cannot rule out the possibility of such a tool being
developed in the future, particularly given the powerful solution techniques
provided by MATLAB. Other areas that may warrant further exploration include:
1. Investigation of the role played by the distribution of V Q ratios in
determining the degree of malalignment
2. Extension of models to include the physiological gases.
3. Mathematical modelling of the change in FRC on induction of
anaesthesia.
I credit my own success in dealing with the problems I faced in this
research to the two years I spent in the Science Faculty at Melbourne University
before transferring to the Medical Faculty. The quality of the lecturers has not
been surpassed. Mentorship also plays an important role – Ian Ritchie and Alex
Robertson – no one could ask for support from more qualified experts. Reaching
out to people one doesn’t know and have never met, can be a daunting prospect.
In my case, I found it a rewarding experience, not only in terms of successful
publications but also in terms of forming long-lasting meaningful relationships
– Bill Mapleson and Ranjan Dash are perfect examples.
41
This would require the simultaneous solution of some 600 equations, 500 of them first order
differential equations, 200 of these non-linear due to the interaction between O2, CO2 and
haemoglobin. A further 300 may become non-linear in the constant inflow case.

The skills required for the investigations undertaken in this thesis are not
possessed by many medical doctors. Neither are they everyone’s cup of tea. With
the conversion of the undergraduate MBBS degree (or its equivalent) to a
postgraduate MD degree in many universities, future medical students have the
opportunity to pursue interests other than those directly relevant to their career
as doctors. I hope that at least some students with an interest in chemistry,
physics and mathematics will take the opportunity to broaden their knowledge
in these areas. One never knows when such skills can be put to good use.

Appendix A: Derivation of distribution
functions for V and Q
The derivation given by Colburn et al. of probability distributions of V
and Q in the case of a lognormal distribution is riddled with typos. The steps in
the derivation are so impenetrable as to possibly render the final result “unsafe”.
We therefore present here a step by step derivation to clarify how the final result
may have been reached. 42
Suppose a lung of total volume V has the property that given any two real
numbers x1 < x2 , the portion of the lung with the log of the ventilation per unit
volume between m1 + σ1 x1 and m1 + σ1 x2 is the same as the portion of the lung
with the log of the blood flow per unit volume between m2 + σ 2 x1 and m2 + σ 2 x2
and that this portion has volume:
V x2
∫
2
dV = e − x /2 dx (A.1)
2π x1
Let = x
y m1 + σ1 x , so that = ( y − m1 ) σ1 and dx = dy σ1 .
When x = x1 , =
y m1 + σ1 x1 ; when x = x2 , =
y m1 + σ1 x2
We may then write Eq. A.1 in the following way:
V m1 +σ1 x2 2
dV = ∫ e −( y − m1 ) /2σ12
dy (A.2)
σ 1 2π m1 +σ1 x1
42
I am indebted to Professor Snezhana Abarzhi for her assistance in tracing the steps used by Colburn
et al. John West was unable to cast any further light on the derivation. Sadly, John Evans passed away
in 2016.
Appendix A: Derivation of distribution functions for V and Q 165

Let the ventilation per unit volume V dV = e y . Then the ventilation of the
portion of lung with volume dV is given by:
V m1 +σ1 x2 2
V = e y ∫ e −( y − m 1 ) /2σ12
dy (A.3)
σ1 2π m1 +σ1 x1
Since e y is a constant it may be taken inside the integration sign.
V m1 +σ1 x2 2
V = ∫ e y e −( y − m 1 ) /2σ12
dy (A.4)
σ 1 2π m1 +σ1 x1
Collecting the indices of the exponential terms:
V m1 +σ1 x2 2
V = ∫ e y −( y − m1 ) /2σ12
dy (A.5)
σ1 2π m1 +σ1 x1
Considering the exponent we have:
2
1  y − m1  1
y−  =
2  σ1 

2σ 12
{2 yσ 12 − ( y − m1 )2 } (A.6)
2
1  y − m1  1
2 {
y−  = −( y 2 − 2m1 y + m12 ) + 2σ 12 y} (A.7)
2  σ1  2σ 1
2
1  y − m1  1
2 {
y−  = − y 2 + 2m1 y − m12 + 2σ 12 y} (A.8)
2  σ1  2σ 1
2
1  y − m1  1
2 {
y−  = − y 2 + 2 y (m1 + σ 12 ) − m12 } (A.9)
2  σ1  2σ 1
166 Appendix A: Derivation of distribution functions for V and Q

2
1  y − m1  1
y− 
2  σ1 
 =
2σ 12
{− y 2 + 2 y(m1 + σ 12 ) − (m1 + σ 12 )2 + (m1 + σ 12 )2 − m12 }
(A.10)
2
1  y − m1 
y− 
2  σ1 
=
1
2σ 1
2 { 2
−  y − (m1 + σ 12 )  + (m1 + σ 12 ) 2 − m12 } (A.11)
2 2
1  y − m1  −  y − (m1 + σ 12 )  (m1 + σ 12 ) 2 − m12
y−  =  + (A.12)
2  σ1  2σ 12 2σ 12
2 2
1  y − m1  −  y − (m1 + σ 12 )  (m1 + σ 12 − m1 )(m1 + σ 12 + m1 )
y−  =  +
2  σ1  2σ 12 2σ 12
(A.13)
2 2
1  y − m1  −  y − (m1 + σ 12 )   σ 12 
y −=   +  m1 + (A.14)
2  σ1  2σ 12  2 
From Eq. A.4,
2
 σ12  −  y − ( m 1+σ12 ) 
 
V m1 +σ1 x2  m1 + 
V =
2  2σ12
σ 1 2π ∫ m1 +σ1 x1
e 
e dy (A.15)
   σ12 
2
−  y − ( m1 +σ12 ) 
 
  V m1 +σ1 x2
2σ12 dy  m 1+ 2 
V = ∫ e
σ 1 
e (A.16)
 2π
m1 +σ1 x1
 
z
Now let = ( y − m1 ) σ1 , =
y m1 + σ1 z , dz = dy σ1 , and substitute in Eq. A.16 to
obtain:
Appendix A: Derivation of distribution functions for V and Q 167

 2 
 V  m 1+ σ21 
2
−[ z −σ1 )]
x2
V =  ∫ e 2
dz  e  
(A.17)
 2π x1 
 
Now let x= z − σ1 , so that:
 σ2
 V x2 −σ1
− x2  m 1+ 21 
V =  ∫ e 2
dx  e  
(A.18)
 2π x1 −σ1 
 
In a similar fashion to Eq. A.17,
 2 
 V  m 2 + σ22 
2
−[ z −σ 2 )]
x2
Q =  ∫ e 2
dz  e  
(A.19)
 2π x1 
 
Let =
a σ 2 − σ1 , so that x − a = z − σ 2 to give the result:
 σ2
 V x2 −σ 2
− ( x − a )2  m 2 + 22 
Q =  ∫ e 2
dx  e  
(A.20)
 2π x1 −σ 2 
 
These are the results given by Colburn et al. (10) in 6 steps not the 20 steps
we have used here. Several questions now remain:
Is Eq. A.1 a realistic depiction of the lung volume or is it merely an abstract
assumption that can be applied to anything in nature?
x2 −σ1
∫
2 2
How do we obtain the result e − x 2 dx = e − x /2
used in deriving the density
x1 −σ1
function for ventilation? Is it for example, by taking the limit as x2 tends to x1 ?
 σ2 
− m 1 + 1
2 
Why does the total ventilation of the lung equal Ve  
which is implied in
the final equation for the density function of ventilation?
168 Appendix A: Derivation of distribution functions for V and Q

Appendix B: Another version of Kety’s Eqs.
When approached, Wolters Kluwer Health Inc., copyright owners of all
material published in ANESTHESIOLOGY refused to allow the reproduction of the
original material used in Figures 2.1 – 2.3, even though the diagrams are freely
available online at the journal site. It was therefore decided that in addition to
pointing the reader to the relevant site, a model of anaesthetic uptake would be
constructed and used to illustrate the point under discussion. Since the
complicated matter of copyright was being investigated while the thesis was
under examination, the material in this appendix was added afterwards.
It was decided to use the correct form of Kety’s equations (42) and extend
the model along the lines of Mapleson’s electric analogues (57, 58). Thus Kety’s
Eq. 2.1 becomes:
dPA 
VL =VIPI − VAPA + λQ t ( P v − Pa ) (B.1)
dt
Assuming the balance gas is insoluble in blood, we have the following mass
balance equations for the balance gas:
(1 − FI )VI =
(1 − FA )VA (B.2)
Since
= (
FI PI PB − PH 2O and
= ) ( )
FA PA PB − PH 2O , Eq. B.2 simplifies to:
VI =
(P − P
B H 2O − PA ) VA (B.3)
(P − P
B H 2O − PI )
With constant outflow, we solve Eq. B.3 for VI , with constant inflow, we solve
Eq. B.3 for VA . Note that we take PB as 760 mmHg and PH 2O as 47 mmHg at 37
Appendix B: Another version of Kety’s Eqs. 169

deg C. The second gas also obeys Eq. B.1 but its partial pressures are so low that
it makes a negligible contribution to the net gas volume change and is therefore
omitted from Eqs. B.3 and B.4. We use inspired concentrations of 0.5% for the
second gas in our simulations.
We next refer to the electronic analogue of Mapleson (1963) which is
based on a standard 70 kg man to extend Kety’s model to a multicompartment
model of the body as suggested by Copperman. Mapleson includes
compartments for the lung, the very rich group of organs in terms of blood supply
(heart, brain, kidneys, liver etc), the muscle group and fat. He does not bother
with the tissues such as tendon and bone which have a poor blood supply.
The effective volume of the lung compartment is given by Mapleson as:
VL =VA + λpV p + λbQb (B.4)
Here VL is the effective lung volume, VA the alveolar air volume, V p the lung
parenchymal tissue volume, Qb the volume of the lung blood pool in equilibrium
with alveolar gas and λp , λb are the parenchymal/gas and blood/gas partition
coefficients, respectively. Mapleson uses the formula of VA = 2.5 + half tidal
volume = 2.75 litre, Vp = 0.6 litre and Qb, = 2 litre so that VL becomes equal to
( 2.75 + 0.6 λp + 2 λb )
The effective volume of the ith tissue compartment is given by:
Vi λiTi + λbQi
= (B.5)
Where Ti is the volume of tissue in the compartment, λi the tissue/gas partition
coefficient and Qi is the volume of blood which is in equilibrium with the tissue.
For the ith tissue compartment, Eq. 2.2 now becomes:
170 Appendix B: Another version of Kety’s Eqs.

dP i
Vi
= λbQ i ( Pa − P i ) (B.6)
dt
Where Pi is the partial pressure of agent in the gas phase with which the tissue is in
equilibrium and Q i is the blood flow in l/min to the tissue. Mapleson suggests the
following values for Ti, Qi, and Q i :
Table B.1. Tissue volume, blood pool and blood flow for various compartments
Compartment Ti Qi Q i
Very rich group (Tv) 6.2 2.53 4.08
Muscle group (Tm) 39.2 0.7 1.07
Fat (Tf) 12.2 0.16 0.26
Mapleson takes the expired alveolar ventilation as 4 l/min and the total blood
flow, obtained from the fourth column, is 5.41 l/min. The tissue/gas partition
coefficients for various agents in each compartment are given in the next table.
They have been taken from various sources including Mapleson (56, 57) and
https://en.wikipedia.org/wiki/Blood%E2%80%93gas_partition_coefficient.
The model does not allow for different circulation times to various organs
and tissues. Neither does it include ventilation-perfusion mismatch. These may
be added at a later date.

Table B.2. Tissue/gas partition coefficients for various agents in several tissues.
Gaseous agent λp λb λv λm λf
nitrous oxide 0.47 0.47 0.47 0.47 1.41
cyclopropane 0.47 0.46 0.46 0.46 7
halothane 6 2.4 6 8 138
isoflurane 3.64 1.4 3.64 4 63
sevoflurane 1.105 0.65 1.105 2.015 31.2
desflurane 0.585 0.45 0.585 0.9 12.15
diethyl ether 12 12 12 12 63
ethylene 0.15 0.14 0.15 0.15 0.45
methoxyflurane 13 13 24 18 800
We next present the MATLAB code which solves the equations. For this I
required much assistance from Ranjan Dash as I had not previously used
MATLAB to solve differential equations. As always, our combined efforts soon
produced the desired result. Note that all green material must be commented out.

function Mapleson_BK_8Apr20
% Written by BK with help from RKD
% Last worked 11Apr20
close all; clear; clc;
% Model parameters
Pb = 760; % Atmospheric barometric pressure (mmHg)
Ph2o = 47; % Pressure of water vapour (mmHg)
% Constant inflow = ci; Constant outflow = co
ci = 1; co = 2;
% Blood flow to each tissue compartment
Q_v = 4.08; % vrg (L/min)
Q_m = 1.07; % muscle (L/min)
Q_f = 0.26; % fat (L/min)
Q_ = [Q_v Q_m Q_f]; % blood flow vector
Q = sum(Q_);% Pulmonary capillary blood flow (l/min)
% Choice of first and second gases

% 1 2 3 4 5 6 7 8
% N2O C3H6 halo iso sevo des ether ethylene
% FG = 7;
% SG = 3;
% Parameters = [Pb Ph2o FG SG];
% Initial conditions used in all but Fig 2.3 where the SG partial pressures initially
equal Fi_SG*(Pb-Ph2o)
Pa0_FG = 0; % Partial pressure of FG in mixed alveolar gas at t = 0
Pvrg0_FG = 0; % Partial pressure of FG in MVB of VRG at t = 0
Pmus0_FG = 0; % Partial pressure of FG in MVB of muscle at t = 0
Pfat0_FG = 0; % Partial pressure of FG in MVB of fat at t = 0
Pa0_SG = 0; % Partial pressure of SG in mixed alveolar gas at t = 0

Pvrg0_SG = 0; % Partial pressure of SG in MVB of VRG at t = 0
Pmus0_SG = 0; % Partial pressure of SG in MVB of muscle at t = 0
Pfat0_SG = 0; % Partial pressure of SG in MVB of fat at t = 0
P0 = [Pa0_FG,Pvrg0_FG,Pmus0_FG,Pfat0_FG,Pa0_SG,Pvrg0_SG,Pmus0_SG,Pfat0_SG]'; %
Initial conditions
options = [];
% Thesis Figure 2.1 - The concentration effect
Fi_FG_vec = [0.01 0.4 0.75 0.90];

Fi_SG = 0.005; % Fraction of inspired SG (halothane)
SG = 3;
tspan = 0:0.01:40;
flow = co;
blue = [173 216 230; 0 191 255; 30 144 255 ;0 0 255]'/255;
red = [250 128 114; 255 0 0; 220 20 60; 255 0 0 ; 128 0 0]'/255;
Fig_21_FG = [1 7];
for j = 1:length(Fig_21_FG)
FG = Fig_21_FG(j);
num_curves = 5 - j;
Parameters = [Pb Ph2o FG SG flow];
for i = 1:num_curves
Fi_FG = Fi_FG_vec(i);
% Solve the ODEs

[t,P] = ode15s(@Mapleson_ODE,tspan,P0,options,Fi_FG,Fi_SG,Parameters);
% State variables
Pa_FG = P(:,1); % Partial pressure of FG in alveoli (mmHg)

Pvrg_FG = P(:,2); % Partial pressure of FG in MVB of VRG (mmHg)
Pmus_FG = P(:,3); % Partial pressure of FG in MVB of muscle (mmHg)
Pfat_FG = P(:,4); % Partial pressure of FG in MVB of fat (mmHg)
%Pv_FG = (Q_v*Pvrg_FG + Q_m*Pmus_FG + Q_f*Pfat_FG)/Q;
% Plot the results

figure(1);
if j == 1
col = blue(:,2)';
else
col = red(:,2)';
end
box off
%pause
plot(t,Pa_FG/(Fi_FG*(Pb-Ph2o)),'Color',col,'LineWidth',1.0); hold on
end
end
xlabel('Time (min)','FontSize',12);
% prepare y-axis label

% Just string together the two pieces needed. It is necessary to write a simple
function to produce the desired string. This passes the text and sizes if doing it
for more than just one or two specific cases.
fmt='\\fontsize{%d}'; % format string for piece
fnLbl=@(FS,S) {[sprintf(fmt,FS,S(1)) sprintf(fmt,FS,S(2)) sprintf(fmt,FS-3,S(3))
sprintf(fmt,FS,'/') sprintf(fmt,FS,S(4)) sprintf(fmt,FS-3,S(5))
sprintf(fmt,FS,S(6))]}
lbl=fnLbl(12,'(FAFI)');
str = ['Fraction of Inspired Concentration ', lbl];
newstr = join(str);
ylabel(newstr,'FontSize',12)
ax = gca;
ax.LineWidth = 1.5;
Font_size = 14;
text(25,0.85,'Nitrous oxide','Interpreter','Tex','Fontsize',Font_size)
text(25,0.50,'Ether','Interpreter','Tex','Fontsize',Font_size)
pause
% Epstein, Rackow, Salanitre and Wolf Fig 1
figure(2);
FG = 1;
SG = 3;
tspan = 0:0.01:5;
flow = co;
Fi_FG_vec = [0.1 0.7];
Fi_SG = 0.005;
for i = 1:length(Fi_FG_vec)
% Solve the ODEs

% State variables
Pa_SG = P(:,5); % Partial pressure of SG in alveoli (mmHg)
Pvrg_SG = P(:,6); % Partial pressure of SG in MVB of VRG (mmHg)
Pmus_SG = P(:,7); % Partial pressure of SG in MVB of muscle (mmHg)
Pfat_SG = P(:,8); % Partial pressure of SG in MVB of fat (mmHg)
Pv_SG = (Q_v*Pvrg_SG + Q_m*Pmus_SG + Q_f*Pfat_SG)/Q;
% Plot the results

plot(t,Pa_SG/(Fi_SG*(Pb-Ph2o)),'-r'); hold on
end
xticks([0 1 2 3 4 5])

xticklabels({'0','1','2','3','4','5'})
ylim([0 0.5])
ylim manual
yticks([0 .1 .2 .3 .4 .5])
yticklabels({'0','.1','.2','.3','.4','.5'})
xlabel('Time (min)');
% Prepare y-axis label
fmt='\\fontsize{%d} %s';% format string for piece
sprintf(fmt,FS,S(6))]};
newstr = join(str);
box off
ax = gca;
ax.LineWidth = 1.5;
Font_size = 9;
s1 = '0.5% halothane with 70% N_2O';
s2 = '0.5% halothane with 10% N_2O';
text(0.5,0.37,s1,'Interpreter','Tex','Fontsize',Font_size,'Rotation',0)
text(3.1,0.26,s2,'Interpreter','Tex','Fontsize',Font_size,'Rotation',0)
pause
% Stoelting and Eger Fig 1

figure(3);
for t =-5:0.1:0
plot(t,1,'k'); hold on
end
FG = 1;
SG_=[8 2 3];
Fi_FG = 0.7;
Fi_SG = 0.005;
flow = co;
tspan = 0:0.01:10;
Kolours = {'b','g','r'};
for j=1:length(SG_)
SG = SG_(j);
% Initial conditions
Pa0_FG = 0; % Partial pressure of FG in mixed alveolar gas at t = 0
Pvrg0_FG = 0; % Partial pressure of FG in MVB of VRG at t = 0
Pmus0_FG = 0; % Partial pressure of FG in MVB of muscle at t = 0
Pfat0_FG = 0; % Partial pressure of FG in MVB of fat at t = 0
Pa0_SG = Fi_SG*(Pb-Ph2o); % Partial pressure of FG in mixed alveolar gas at t =
0
Pvrg0_SG = Fi_SG*(Pb-Ph2o); % Partial pressure of FG in MVB of VRG at t = 0
Pmus0_SG = Fi_SG*(Pb-Ph2o); % Partial pressure of FG in MVB of muscle at t = 0
Pfat0_SG = Fi_SG*(Pb-Ph2o); % Partial pressure of FG in MVB of fat at t = 0
P0 = [Pa0_FG,Pvrg0_FG,Pmus0_FG,Pfat0_FG,Pa0_SG,Pvrg0_SG,Pmus0_SG,Pfat0_SG]';
% Solve the ODEs

% State variables
Pa_SG = P(:,5); % Partial pressure of SG in alveoli (mmHg)
Pvrg_SG = P(:,6); % Partial pressure of SG in MVB of VRG (mmHg)
Pmus_SG = P(:,7); % Partial pressure of SG in MVB of muscle (mmHg)
Pfat_SG = P(:,8); % Partial pressure of SG in MVB of fat (mmHg)
colour = Kolours{j}
plot(t,(Pa_SG/(Fi_SG*(Pb-Ph2o))),colour,'LineWidth',2); hold on
end

xlabel('Time (min)');
fmt='\\fontsize{%d} %s'; % format string for piece
newstr = join(str);
ylabel(newstr)
box off
ax = gca;
ax.LineWidth = 1.5;
xticks([-5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10])
xticklabels({'-5','-4','-3','-2','-1','0','1','2','3','4','5','6','7','8','9','10'})
Font_size = 12;
s1 = ['70% N_2O' newline 'Added'];
text(-4,1.2,s1,'Interpreter','Tex','Fontsize',Font_size,'Rotation',0)
annotation('arrow',[4.2 5.7]/15,[1.6 0.59]/5,'LineWidth',4);
pause
% Figure 4.1 in thesis
figure(4)
% Commonly used inspired concentrations of N2O, des, sevo, iso and halo as first gas
FG_vec = [1 6 5 4 3]; % identifies first gas so appropriate
inspired concentration is chosen for calculations in ode15s
Fi_FG_vec = [0.75 0.06 0.01 0.015 .01]; % Lists the FG concentration commonly
used
Fi_SG = 0.005; % Fraction of inspired SG included solely to satisfy equations. We are

only interested in the FG here. If no SG, the Eqs won’t work. We use halothane but it
is immaterial.
SG = 3;
tspan = 0:0.01:30;
flow = co;
Kolours = ['r' 'b' 'y' 'c' 'g'];
num_curves = length(Fi_FG_vec);
for i = 1:num_curves
FG = FG_vec(i);
% Solve the ODEs

% State variables
Pa_FG = P(:,1); % Partial pressure of FG in alveoli (mmHg)
Pvrg_FG = P(:,2); % Partial pressure of FG in MVB of VRG (mmHg)
Pmus_FG = P(:,3); % Partial pressure of FG in MVB of muscle (mmHg)
Pfat_FG = P(:,4); % Partial pressure of FG in MVB of fat (mmHg)
col = Kolours(i);
plot(t,Pa_FG/(Fi_FG*(Pb-Ph2o)),'Color',col,'LineWidth',1.0);
hold on
pause
end
box off
xlabel('Time (min)','FontSize',12);

fmt='\\fontsize{%d} %s'; % format string for piece

newstr = join(str);
ax = gca;
ax.LineWidth = 1.5;
Font_size = 9;
text(22.5,1.0,'Nitrous oxide','Interpreter','Tex','Fontsize',Font_size)
text(22.5,0.92,'Desflurane','Interpreter','Tex','Fontsize',Font_size)
text(22.5,0.80,'Sevoflurane','Interpreter','Tex','Fontsize',Font_size)
text(22.5,0.63,'Isoflurane','Interpreter','Tex','Fontsize',Font_size)
text(22.5,0.50,'Halothane','Interpreter','Tex','Fontsize',Font_size)
function dP = Mapleson_ODE(~,P,Fi_FG,Fi_SG,Parameters)
% This function solves the first order differential equations
% The vectors VL_, VT_,BP_, and Q_ are shown but only Q_ is actually used
ci = 1; co = 2;
% Unpack the entries in Parameters
Pb = Parameters(1);
Ph2o = Parameters(2);
FG = Parameters(3);
SG = Parameters(4);
flow = Parameters(5);
%pause
% Lung volumes
VL_a = 2.75; % Volume of alveolar air (litre)
VL_p = 0.6; % Volume of parenchymal tissue (litre)
VL_b = 2; % Volume of blood in equilibrium with lung compartment (litre)
%VL_ = [VL_a VL_p VL_b]; % Volume vector
% Tissue volumes
VT_v = 6.2; % very rich group (litre)
VT_m = 39.2; % muscle group (litre)
VT_f = 12.2; % fat (litre)
%VT_ = [VT_v VT_m VT_f]; % Volume vector
% Blood pool in equilibrium with tissue compartment

q_v = 2.53; % vrg
q_m = 0.7; % muscle
q_f = 0.16; % fat
%BP_ = [q_v q_m q_f]; % Blood pool vector
% Blood flow to each tissue compartment

Q_v = 4.08; % vrg (L/min)
Q_m = 1.07; % muscle (L/min)
Q_f = 0.26; % fat (L/min)
Q_ = [Q_v Q_m Q_f]; % blood flow vector
% Tissue/gas partition coefficients

% N2O C3H6 hal iso sevo des ether ethylene
Lp = [0.47 0.46 6 3.64 1.105 0.585 12 0.15]; % lung parenchyma/gas
Lb = [0.47 0.47 2.4 1.4 0.65 0.45 12 0.14]; % blood/gas
Lv = [0.47 0.46 6 3.64 1.105 0.585 12 0.15]; % VRG/gas
Lm = [0.47 0.46 8 4 2.015 0.9 12 0.15]; % muscle/gas
Lf = [1.41 7 138 63 31.2 12.15 63 0.45]; % fat/gas
L = [Lp' Lb' Lv' Lm' Lf'];
if flow == co
Ve = 4.0; % Expired alveolar ventilation (l/min)
else
Vi = 4.0; % Inspired alveolar ventilation
end
Q = sum(Q_);% Pulmonary capillary blood flow (l/min)
% State variables

Pa_FG = P(1); % Partial pressure of FG in alveoli (mmHg)
Pvrg_FG = P(2); % Partial pressure of FG in MVB of VRG (mmHg)
Pmus_FG = P(3); % Partial pressure of FG in MVB of muscle (mmHg)
Pfat_FG = P(4); % Partial pressure of FG in MVB of fat (mmHg)
Pv_FG = (Q_v*Pvrg_FG + Q_m*Pmus_FG + Q_f*Pfat_FG)/Q;
Pa_SG = P(5); % Partial pressure of SG in alveoli (mmHg)

Pvrg_SG = P(6); % Partial pressure of SG in MVB of VRG (mmHg)
Pmus_SG = P(7); % Partial pressure of SG in MVB of muscle (mmHg)
Pfat_SG = P(8); % Partial pressure of SG in MVB of fat (mmHg)
Pi_FG = Fi_FG*(Pb-Ph2o); % Partial pressure of inspired FG

Pi_SG = Fi_SG*(Pb-Ph2o); % Partial pressure of inspired SG
if flow == co
Vi = (Pb-Ph2o-Pa_FG)*Ve/(Pb-Ph2o-Pi_FG); % Inspired ventilation (L/min)
else
Ve = (Pb-Ph2o-Pi_FG)*Vi/(Pb-Ph2o-Pa_FG); % Expired ventilation (L/min)
end
VL = VL_a + L(FG,1)*VL_p + L(FG,2)*VL_b;

dPa_FG = (Vi*Pi_FG - Ve*Pa_FG + L(FG,2)*Q*(Pv_FG-Pa_FG))/VL;
dPvrg_FG = (L(FG,2)*Q_v*(Pa_FG - Pvrg_FG))/(L(FG,3)*VT_v + L(FG,2)*q_v);
dPmus_FG = (L(FG,2)*Q_m*(Pa_FG - Pmus_FG))/(L(FG,4)*VT_m + L(FG,2)*q_m);
dPfat_FG = (L(FG,2)*Q_f*(Pa_FG - Pfat_FG))/(L(FG,5)*VT_f + L(FG,2)*q_f);
%dPv_FG =(Q_v*dPvrg_FG + Q_m*dPmus_FG + Q_f*dPfat_FG)/Q;
VL = VL_a + L(SG,1)*VL_p + L(SG,2)*VL_b;

dPa_SG = (Vi*Pi_SG - Ve*Pa_SG + L(SG,2)*Q*(Pv_SG-Pa_SG))/VL;
dPvrg_SG = (L(SG,2)*Q_v*(Pa_SG - Pvrg_SG))/(L(SG,3)*VT_v + L(SG,2)*q_v);
dPmus_SG = (L(SG,2)*Q_m*(Pa_SG - Pmus_SG))/(L(SG,4)*VT_m + L(SG,2)*q_m);
dPfat_SG = (L(SG,2)*Q_f*(Pa_SG - Pfat_SG))/(L(SG,5)*VT_f + L(SG,2)*q_f);
%dPv_SG =(Q_v*dPvrg_SG + Q_m*dPmus_SG + Q_f*dPfat_SG)/Q;
dP = [dPa_FG,dPvrg_FG,dPmus_FG,dPfat_FG,dPa_SG,dPvrg_SG,dPmus_SG,dPfat_SG]';
% dP contains the solutions to all 8 differential equations.

Appendix C: Sample MATLAB routines
The computer programs used to produce Figure 5.1 and Figure 8.1 were
written in Microsoft Visual Basic. It was only after running into problems with
these programs that I allowed myself to be convinced that a switch to MATLAB
would be beneficial. The power and scope of this language soon became evident.
I will be forever grateful to Ranjan Dash for insisting that I learn MATLAB if I
wished to continue receiving his help in debugging faulty code. It is not practical
to include the code of every program written to produce the results used in this
thesis. Instead, a selection of these programs will be presented. Note that all
wording shown in green is to be commented out.
APPENDIX C.1: ROUTINES USED TO DRAW FIGURES 6.1, 6.2 AND 6.3
% Pp_solubility_dV_main
% This program furnishes the partial pressure-solubility diagram for
the second gas in the form of PG2/Pi2. Values are obtained for mixed
alveolar gas and mixed pulmonary end-capillary blood.
% A log normal distribution of ventilation and blood flow is
assumed, having a degree of mismatch given by sigma.
% WE STUDY THE EFFECT OF CHANGING (Vi - Va) I.E. WE DON'T FIX FiN2O
% The volume change dV = Vi - Va is kept constant as sigma is
varied.
% This means that FiN2O is changed to keep dV constant.
% The values of FiN2O are determined by the program
FiN2O_Fsolve_main.m found in the Fsolve_FiN2O directory.
An equivalent function is performed below by the function
Call_Fsolve_FiFG_25Mar19.m
% Input variables are:

% L1, solubility of the first gas
% L2, solubility of the second gas
% Fi1,Pi1, inspired concentration and partial pressure of first gas
% Fi2,Pi2, inspired concentration and partial pressure of second gas
% Fv1,Pv1, concentration and partial pressure of first gas in
% mixed venous blood
% Fv2, Pv2, concentration and partial pressure of second gas in
% mixed venous blood( = 0 )
% V0, total expired alveolar ventilation
% Q0, total pulmonary blood flow
% sigma, degree of V/Q mismatch
Appendix C: Sample MATLAB routines 179

% Output variables are:
% PA/Pi for the second gas
% Pa/Pi for the second gas
% Vi, Vi-Va
% Written by Dr Ben Korman BSc, MBBS, MD, FANZCA

% Started 13/11/2017
% Last revised 25Sep19
close all; clear all; clc;
global Pb PH2O
Pb = 760; % Barometric pressure of the air, mmHg
PH2O = 47; % Partial pressure of H2O in the
inspired air, mmHg (T=37oC)
L1 = 0.47; % Solubility of N2O
% Set up vector L of L2 values

L = [0.00001:0.00001:0.0001, 0.0001:0.0001:0.001, 0.001:0.001:0.090,
0.09:0.01:1.0, 1:.1:15, 15:1:100, 100:10:1000, 1000:100:10000,
10000:1000:100000];
num_sols = length(L);
FiN2O = 0.7; % Inspired concn of N2O
Fv1 = 0;
Fv2 = 0;
V0 = 4;
Q0 = 5;
dt = 0.01;
N = 2*(1/dt-1); %Partition used in applying trapezoidal rule.
% Previously determined values of inspired FiN2O required to achieve

% the value of (Vi-Va) shown as a comment on the right
% for values of sigma shown on the next line.
% TABLE 1
% sigma 0 0.5 1 1.5 2 dV
FiN2O_(1,:) = [0.0334 0.0353 0.0413 0.0440 0.0712]; % 0.05
FiN2O_(2,:) = [0.0659 0.0696 0.0812 0.1025 0.1379]; % 0.10
FiN2O_(3,:) = [0.0977 0.1030 0.1198 0.1504 0.2004]; % 0.15
FiN2O_(4,:) = [0.1287 0.1356 0.1571 0.1961 0.2588]; % 0.20
FiN2O_(5,:) = [0.1589 0.1673 0.1932 0.2397 0.3134]; % 0.25
FiN2O_(6,:) = [0.1885 0.1982 0.2280 0.2814 0.3645]; % 0.30
FiN2O_(7,:) = [0.2174 0.2283 0.2618 0.3212 0.4123]; % 0.35
FiN2O_(8,:) = [0.2456 0.2576 0.2944 0.3592 0.4570]; % 0.40
FiN2O_(9,:) = [0.2732 0.2862 0.3260 0.3955 0.4987]; % 0.45
FiN2O_(10,:)= [0.3002 0.3141 0.3566 0.4302 0.5376]; % 0.50
FiN2O_(11,:)= [0.3266 0.3413 0.3862 0.4633 0.5740]; % 0.55
FiN2O_(12,:)= [0.3525 0.3679 0.4149 0.4949 0.6079]; % 0.60
FiN2O_(13,:)= [0.3777 0.3937 0.4426 0.5251 0.6395]; % 0.65
FiN2O_(14,:)= [0.4024 0.4190 0.4694 0.5539 0.6690]; % 0.70
FiN2O_(15,:)= [0.4267 0.4437 0.4954 0.5814 0.6964]; % 0.75
FiN2O_(16,:)= [0.4504 0.4678 0.5206 0.6076 0.7220]; % 0.80
% Set up "Other" array
180 Appendix C: Sample MATLAB routines

Other = {L,L1,dt,V0,Q0, FiN2O_,Fv1};
% First we obtain values for PA/Pi with sigma = 0 and dV = 0 mL/min

% NB when sigma = 0, PA_Pi = Pa_Pi, as we have a homogeneous lung.
DV = 0;
sigma = 0;
[PA2_Pi2,Pa2_Pi2] = Get_PAPi_PaPi(DV,sigma,Other);
PA2_Pi2_0_0 = PA2_Pi2;
Pa2_Pi2_0_0 = Pa2_Pi2;
% We could just calculate PA2_Pi2 and Pa2_Pi2 values for each value
of sigma and dV at the start and just use them when necessary in
drawing the respective graphs but the Get_PAPi routine is so fast
that I often just call it again and again.
% Now graph these results for Alveolar gas in Figure 6.1A.

num_graph = 1;
figure(num_graph);
set(figure(num_graph),'Units','inches','Position',[0 0.3 20 10]);
xlim([-4 4])
ylim([0 1.5])
xlabel('log_{10}(\lambda)')
ylabel({'P_A/P_I ';' ';'or ';' ';'P_a/P_I
'}15
6.140
)
hYLabel = get(gca,'YLabel');
set(hYLabel,'rotation',0,'VerticalAlignment','middle')
set(gca,'Fontsize',16,'Linewidth',1.5,'box','on','FontName', 'Times
New Roman', 'FontWeight', 'bold')
% Plot vertical lines at sigma = log(0.007) and log(13)

y1 = get(gca,'ylim');
idx = log10(0.007);
hold on;
plot([idx idx], y1,'g','LineStyle', ':','Linewidth',2)
idx = log10(13);
hold on
% Next we obtain values for PA/Pi and Pa/Pi with dV = 0 and sigma =
0 and 1.5
DV_choice = 0 / 1000;
DV = DV_choice;
sigma = 1.5;
PA2_Pi2_15_0 = PA2_Pi2;
Pa2_Pi2_15_0 = Pa2_Pi2;
% Now graph these results for Alveolar gas in Figure 1

plot(log10(L),PA2_Pi2_0_0,'k','LineStyle', '-.','Linewidth',3)
plot(log10(L),PA2_Pi2_15_0,'b',log10(L),Pa2_Pi2_15_0,'r','LineStyle'
, '-.','Linewidth',3)
txt = 'A';
text(3.0,1.25,txt,'FontSize',40,'FontName','Times New Roman')

hold on
pause
% Now graph these results for Alveolar gas in Figure 6.1B.

num_graph = 2;
figure(num_graph);
xlim([-4 4])
ylim([0 1.5])

idx = log10(0.007);
hold on;
idx = log10(13);
hold on
'})
num_fig = num2str(num_graph);
% Next we obtain values for PA/Pi and Pa/Pi with sigma = 0 and dV =
800 mL/min
DV = 600 / 1000;
sigma = 0;
[PA2_Pi2,~] = Get_PAPi_PaPi(DV,sigma,Other);
PA2_Pi2_0_600 = PA2_Pi2;

plot(log10(L),PA2_Pi2_0_600,'k','LineStyle', '-','Linewidth',3)
txt = 'B';
hold on
pause
%*******************************************
% Next we obtain Fig. 6.1C by adding the results in Fig. 6.1A and
6.1B
num_graph = 3;
figure(num_graph);
xlim([-4 4])
ylim([0 1.5])
'})


idx = log10(0.007);
hold on;
idx = log10(13);
hold on
% Next we obtain values for PA/Pi and Pa/Pi with dV = 0 and sigma =
0 and 1.5
PA2_PI2_15_0 = PA2_Pi2_15_0+PA2_Pi2_0_600-PA2_Pi2_0_0;
Pa2_PI2_15_0 = Pa2_Pi2_15_0+PA2_Pi2_0_600-PA2_Pi2_0_0;

plot(log10(L),PA2_PI2_15_0,'b',log10(L),Pa2_PI2_15_0,'r','LineStyle'
, '-.','Linewidth',3)
txt = 'C';
hold on
pause
%**************************************************************
% Graphs for Fig. 6.2A and 6.2B
num_graph = 4;
figure(num_graph);
xlim([-4 4])
ylim([0 1.5])

idx = log10(0.007);
hold on;
idx = log10(13);
hold on
ylabel({'P_A/P_I '})
800 mL/min

DV = 600 / 1000;
sigma = 0;
PA2_Pi2_0_600 = PA2_Pi2;

txt = 'A';
hold on
DV = 0;
sigma = 1.5;
PA2_Pi2_15_0 = PA2_Pi2;

plot(log10(L),PA2_Pi2_15_0,'b','LineStyle', '-.','Linewidth',3)
hold on
% Next we obtain PA/Pi and Pa/Pi with sigma = 1.5 and dV = 600
mL/min
DV = 600/1000;
PA2_Pi2_15_800 = PA2_Pi2;

plot(log10(L),PA2_Pi2_15_800,'b','LineStyle', '-','Linewidth',3)
hold on
pause
% Now graph the results for Arterial blood in Figure 6.2B

num_graph = 5;
figure(num_graph);
figure(num_graph);
xlim([-4 4])
ylim([0 1.5])

idx = log10(0.007);
hold on;
idx = log10(13);
hold on
ylabel({'P_a/P_I '})

800 mL/min
DV = 600 / 1000;
sigma = 0;
[~,Pa2_Pi2] = Get_PAPi_PaPi(DV,sigma,Other);
Pa2_Pi2_0_600 = Pa2_Pi2;

txt = 'B';
hold on
DV = 0;
sigma = 1.5;
Pa2_Pi2_15_0 = Pa2_Pi2;

plot(log10(L),Pa2_Pi2_15_0,'r','LineStyle', '-.','Linewidth',3)
hold on
mL/min
DV = 600/1000;
Pa2_Pi2_15_600 = Pa2_Pi2;

plot(log10(L),Pa2_Pi2_15_600,'r','LineStyle', '-','Linewidth',3)
hold on
pause
%*************************************
% AUGMENTATION RATIOS
% We plot ARg and ARb for values of (Vi-Va) = 200 and 600 mL/min
% with sigma = 1.5.
% As we already have PA/Pi and Pa/Pi wnen sigma = 0 and sigma = 1.5
% for dV = 0 and dV = 600 mL/min, we only need to determine values
for sigma = 1.5 and dV = 200 mL/min
mL/min
DV = 200 / 1000;
sigma = 1.5;
PA2_Pi2_15_200 = PA2_Pi2;
Pa2_Pi2_15_200 = Pa2_Pi2;

mL/min
DV = 600 / 1000;
sigma = 1.5;
PA2_Pi2_15_600 = PA2_Pi2;
Pa2_Pi2_15_600 = Pa2_Pi2;
% GRAPHS OF AR(gas) and AR(blood) for sigma = 1.5 with (Vi-Va) =

0.2,
% and 0.8 L/min
num_graph = 6;
figure(num_graph);
sigma = 1.5;
ARg = PA2_Pi2_15_200./PA2_Pi2_15_0;
ARb = Pa2_Pi2_15_200./Pa2_Pi2_15_0;
plot(log10(L),ARg,'b',log10(L),ARb,'r','LineStyle','-.',
'Linewidth',3)
hold on
xlim([-4 4])
ylim([1 1.6])
xlabel('log10(\lambda)','FontName','Times New Roman','FontSize',40)
ylabel('Augmentation ratio','FontName','Times New
Roman','FontSize',40)
set(gca,'Fontsize',20,'Linewidth',1.5,'box','on')
idx = log10(0.007);
hold on
idx = log10(13);
hold on
ARg = PA2_Pi2_15_600./PA2_Pi2_15_0;
ARb = Pa2_Pi2_15_600./Pa2_Pi2_15_0;
plot(log10(L),ARg,'b',log10(L),ARb,'r','LineStyle','-',
'Linewidth',3)
hold on
txt = 'A';
hold on
pause
%***************************************
% Now plot ARg and ARb for (Vi-Va) = 800 mL/min for various values
of sigma
% This is Fig. 6.3B
num_graph = 7;
figure(num_graph);
% plot(log10(L),ARg,'b',log10(L),ARb,'r','LineStyle','-',
'Linewidth',3)
% hold on
xlim([-4 4])

ylim([0.8 2.4])
xlabel('log10(\lambda)','FontName','Times New Roman','FontSize',40)
ylabel('Augmentation ratio','FontName','Times New
Roman','FontSize',40)
set(gca,'Fontsize',20,'Linewidth',1.5,'box','on')
hold on
idx = log10(0.007);
idx = log10(13);
hold on
sigma = 0.5;
DV = 0;
PA2_Pi2_5_0 = PA2_Pi2;
Pa2_Pi2_5_0 = Pa2_Pi2;
DV = 800 / 1000;
PA2_Pi2_5_800 = PA2_Pi2;
Pa2_Pi2_5_800 = Pa2_Pi2;
ARg = PA2_Pi2_5_800./PA2_Pi2_5_0;
ARb = Pa2_Pi2_5_800./Pa2_Pi2_5_0;
'Linewidth',3)
hold on
pause
sigma = 1;
DV = 0;
PA2_Pi2_10_0 = PA2_Pi2;
Pa2_Pi2_10_0 = Pa2_Pi2;
DV = 800 / 1000;
PA2_Pi2_10_800 = PA2_Pi2;
Pa2_Pi2_10_800 = Pa2_Pi2;
ARg = PA2_Pi2_10_800./PA2_Pi2_10_0;
ARb = Pa2_Pi2_10_800./Pa2_Pi2_10_0;
'Linewidth',3)
hold on
pause
sigma = 1.5;
DV = 0;
PA2_Pi2_15_0 = PA2_Pi2;
Pa2_Pi2_15_0 = Pa2_Pi2;
DV = 800 / 1000;
PA2_Pi2_15_800 = PA2_Pi2;
Pa2_Pi2_15_800 = Pa2_Pi2;
ARg = PA2_Pi2_15_800./PA2_Pi2_15_0;
ARb = Pa2_Pi2_15_800./Pa2_Pi2_15_0;
'Linewidth',3)

hold on
pause
sigma = 2;
DV = 0;
PA2_Pi2_20_0 = PA2_Pi2;
Pa2_Pi2_20_0 = Pa2_Pi2;
DV = 800 / 1000;
PA2_Pi2_20_800 = PA2_Pi2;
Pa2_Pi2_20_800 = Pa2_Pi2;
ARg = PA2_Pi2_20_800./PA2_Pi2_20_0;
ARb = Pa2_Pi2_20_800./Pa2_Pi2_20_0;
'Linewidth',3)
hold on
pause
%Draw horizontal black line at sigma = 0, dV = 0

x1 = get(gca,'xlim');
idy = 1.2;
plot(x1, [idy idy],'k','LineStyle', '-','Linewidth',1.5)
hold on
pause
x = [0.22 0.5];
y = [0.33 0.7];
annotation('textarrow',x,y,'LineWidth',2)
hold on
x = [0.8 0.6];
y = [0.3 0.2];
annotation('textarrow',x,y,'LineWidth',2)
hold on
txt = 'B';
hold on
%***************************
return
function [PA2_Pi2,Pa2_Pi2] = Get_PAPi_PaPi(DV,sigma,Other)

% Function that returns the vectors for PA/Pi and Pa/Pi
% Written by Ben Korman
% Started on 10/12/2017
% sigma value 0 0.5 1.0 1.5 2.0 volume uptake

DV_val L/min
%FiN2O_(0,:) = [ 0 0 0 0 0 % 0.00
% FiN2O_(1,:) = [0.0334 0.0353 0.0413 0.0440 0.0712]; % 0.05
% FiN2O_(2,:) = [0.0659 0.0696 0.0812 0.1025 0.1379]; % 0.10
% FiN2O_(3,:) = [0.0977 0.1030 0.1198 0.1504 0.2004]; % 0.15
% FiN2O_(4,:) = [0.1287 0.1356 0.1571 0.1961 0.2588]; % 0.20
% FiN2O_(5,:) = [0.1589 0.1673 0.1932 0.2397 0.3134]; % 0.25
% FiN2O_(6,:) = [0.1885 0.1982 0.2280 0.2814 0.3645]; % 0.30

% FiN2O_(7,:) = [0.2174 0.2283 0.2618 0.3212 0.4123]; % 0.35
% FiN2O_(8,:) = [0.2456 0.2576 0.2944 0.3592 0.4570]; % 0.40
% FiN2O_(9,:) = [0.2732 0.2862 0.3260 0.3955 0.4987]; % 0.45
% FiN2O_(10,:)= [0.3002 0.3141 0.3566 0.4302 0.5376]; % 0.50
% FiN2O_(11,:)= [0.3266 0.3413 0.3862 0.4633 0.5740]; % 0.55
% FiN2O_(12,:)= [0.3525 0.3679 0.4149 0.4949 0.6079]; % 0.60
% FiN2O_(13,:)= [0.3777 0.3937 0.4426 0.5251 0.6395]; % 0.65
% FiN2O_(14,:)= [0.4024 0.4190 0.4694 0.5539 0.6690]; % 0.70
% FiN2O_(15,:)= [0.4267 0.4437 0.4954 0.5814 0.6964]; % 0.75
% FiN2O_(16,:)= [0.4504 0.4678 0.5206 0.6076 0.7220]; % 0.80
%DV_pos = [ 0 1 2 3 4 5 6 7 8 9 10 11
12 13 14 15 16 ];
%DV_val = [0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.4 0.45 0.50
0.55 0.60 0.65 0.70 0.75 0.80];
%sig_pos = [ 0 1 2 3 4 ];
%sig_val = [0.0 0.5 1.0 1.5 2.0 ];
%Other = {L,L1,dt,V0,Q0,DV_val,sig_val,FiN2O_,Fv1}
L = Other{1}
L1 = Other{2};
dt = Other{3};
V0 = Other{4};
Q0 = Other{5};
FiN2O_ = Other{6};
Fv1 = Other{7};
num_sols = length(L);
N = 2*(1/dt-1);
sig_pos = 2*sigma + 1;
DV_pos = DV*20;
if DV == 0
Fi1 = 0;
else
Fi1 = FiN2O_(DV_pos,sig_pos);
end
PA2_Pi2 = zeros(1,num_sols);
Pa2_Pi2 = zeros(1,num_sols);
for j = 1:num_sols
L2 = L(j);
t = zeros(1,N+1);
PA_Pi = zeros(1,N+1);
Pa_Pi = zeros(1,N+1);
for i = 1:N+1
t(i) = -1 + i*dt;
ti = t(i);
x = ti/(1-ti*ti);
dx = dt*(1+ti^2)/(1-ti^2)^2;
f_e = (V0/Q0)*exp(-(2*sigma*x-sigma*sigma)/2);
f_x = ((f_e + L1*(1-Fv1))/(f_e + L1*(1-Fi1)))*f_e/(f_e +
L2);
PA_Pi(i) = (exp(-x*x/2)) * f_x * dx;
Pa_Pi(i) = (exp(-(x-sigma)*(x-sigma)/2)) * f_x * dx;
end

PA2_Pi2(j) = (sum(PA_Pi) -
(PA_Pi(1)+PA_Pi(length(PA_Pi))/2))/sqrt(2*pi);
Pa2_Pi2(j) = (sum(Pa_Pi) -
(Pa_Pi(1)+Pa_Pi(length(Pa_Pi))/2))/sqrt(2*pi);
end
APPENDIX C.2 ROUTINES USED TO DRAW FIGURE 7.4 & FIGURE 7.5
% Main program of routine to draw Figs 8A and 8B
close all; clear all; clc

format long
V0 = 4;
Q0 = 5;
V0_Q0 = V0/Q0;
s = 2; % symbol for sigma
% Enter parameters for FG

L1 = 0.47; % Solubility of first gas
Fv = 0; % Fractional concn of gas 1 in gas sample in eqlm with
venous blood
Fi = 0.7; % Fractional concn of gas 1 in fresh gas mixture
F2i = 0.01; % Fractional concn of gas 2 in fresh gas mixture
% Determine volume uptake = 757 ml

fun = @(x) 1./((L1*(1-Fi)*(exp(((x-
s/2).^2)/2)))+(V0_Q0*(exp(((x+s/2).^2)/2))));
dV = integral(fun,-Inf,Inf);
dV = L1*Fi*V0*dV/sqrt(2*pi)
sigma = 0:0.01:2;
num_sigmas = length(sigma);
L2_max_xu = zeros(1,num_sigmas);
Max_L2_value = zeros(1,num_sigmas);
L2vec = [0.0001:0.00001:0.001 0.001:0.0001:0.01 0.01:0.001:0.1

0.1:0.001:1 1:0.01:10 10:.1:100 100:1:1000]; %This gives good
coverage
num_L2 = length(L2vec);
G2u_12 = zeros(num_sigmas,num_L2); % SGE for each value of sigma and

L2
s_max = zeros(num_sigmas,num_L2);
Correct_AR = zeros(num_sigmas, num_L2);
s_max_xu = zeros(1, num_sigmas);
Max_s_value = zeros(1, num_sigmas);
for k = 1:num_sigmas
%% Set the value of s being investigated
s = sigma(k);
VarVec = [V0 Q0 s L1 Fv];
Guess = 0.5;

[Eqs,F1i] = Call_Fsolve_FiFG_25Mar19(Guess,dV,VarVec)
if real(F1i) > 1
Eqs
F1i
return
end
L1p = L1*(1-F1i);
F1v = Fv;
dx = 0.01;
x = -4+s/2:dx:4+s/2;
num_points = length(x);
% Now allocate memory for all variables changing with x for this
value of sigma
Va = zeros(1,num_points);
V1i = zeros(1,num_points);
V1u = zeros(1,num_points);
V1r = zeros(1,num_points);
Vi = zeros(1,num_points);
Qi = zeros(1,num_points);
V2u_01 = zeros(1,num_points);
V2r_01 = zeros(1,num_points);
for i = 1:num_points
z = x(i);
% Work on Vi first for this value of sigma and V1
Va(i) = (V0/sqrt(2*pi))*exp((-(z-s/2)^2)/2);
V1u(i) = 1/((L1p*exp(((z-
s/2)^2)/2))+V0_Q0*exp(((z+s/2)^2)/2));
V1u(i) = (L1*F1i*V0/sqrt(2*pi))*V1u(i); % uptake of FG
Vi(i) = Va(i)+V1u(i);
V1i(i) = Vi(i)*F1i; % inflow of FG
V1r(i) = V1i(i) - V1u(i); % remainder of FG after uptake
Qi(i) = (Q0/sqrt(2*pi))*exp((-(z+s/2)^2)/2);
end
% With sigma specified, V1u is determined, so now we calculate

V2u for the first and second steps for each value of L2
for j = 1:num_L2
L2 = L2vec(j);
V2i(i) = Vi(i)*F2i;
V2u_01(i) = L2*F2i*Vi(i)/(L2 + Vi(i)/Qi(i));
V2r_01(i) = V2i(i) - V2u_01(i);
% Calculate Total uptake of SG

V2u_02(i) = L2*F2i*Vi(i)/(L2 + Va(i)/Qi(i));
% Calculate uptake of SG in step2 by difference
V2u_12(i) = V2u_02(i)-V2u_01(i);
end
SGE = sum(V2u_12)-(V2u_12(1)+V2u_12(num_points))/2;
SGE = SGE * dx;
G2u_12(k,j) = SGE;

Step1_uptake = sum(V2u_01)-(V2u_01(1)+V2u_01(num_points))/2;
Step1_uptake = Step1_uptake * dx;
Fractional_SGE = SGE/Step1_uptake;
Correct_AR(k,j) = Fractional_SGE;
end
G2u_12;
[Max_L2xu, I] = max(G2u_12(k,:).');
Max_Uptake_soly = L2vec(I);
L2_max_xu(k) = Max_L2xu;
Max_L2_value(k) = Max_Uptake_soly;
Correct_AR;
end
num_fig = 1;
figure(num_fig); set(figure(num_fig),'Units','inches','Position',[0
0.5 20 10]);
ylim([0 .0025])
xlim([-4 3])
xlabel('Solubility of second gas in blood, log_{10}(\lambda_{SG)')
ylabel('Second gas uptake in step 2 (l/min x 10^3)')
Title = 'Second gas uptake in step 2 as a function of
log_{10}(\lambda_{SG})';
title(Title)
hold on
%Draw a vertical line at x = 0
idx = 0;
plot([idx idx], y1,'k','LineStyle','-')
hold on;
plot(log10(L2vec),G2u_12(num_sigmas,:),'b','LineWidth',4)
hold on
pause
num_fig = 2;
0.5 20 10]);
ylim([0 .0025])
xlim([-4 3])
xlabel('Solubility of second gas in blood, log_{10}(\lambda_{SG})')
ylabel('Second gas uptake in step 2 (l/min x 10^3)')
Title = 'Second gas uptake in step 2 as a function of
log_{10}(\lambda_{SG}) for various V/Q various V/Q values';
title(Title)
hold on
idx = 0;
plot([idx idx], y1,'k','LineStyle','-')
hold on;
k = 1;
plot(log10(L2vec),G2u_12(k,:),'Color',[1,0.2,0],'LineWidth',10)
hold on
for k = 2:num_sigmas-1
if k == 1
plot(log10(L2vec),G2u_12(k,:),'p','LineWidth',10)
elseif k == num_sigmas

plot(log10(L2vec),G2u_12(k,:),'b','LineWidth',4)
else
plot(log10(L2vec),G2u_12(k,:),'Color',[0.9,1,0],'LineWidth',1)
end
hold on
end
k = num_sigmas;
plot(log10(L2vec),G2u_12(k,:),'b','LineWidth',4)
hold on
% Draw a horizontal line for Alignment

y1 = 2.1*10^-3;
idy = [y1 y1];
plot([-4 -0.23], idy,'k','LineStyle','-','LineWidth', 4)
hold on;
txt = 'ALIGNMENT';
text(-3.0,2.2*10^-3,txt,'FontSize',35)
hold on
% Add letter A
txt = 'A';
text(-0.6,1.9*10^-3,txt,'Fontsize',30)
hold on
% Draw a horizontal line for Retention

idy = [0.75*10^-3 0.75*10^-3];
plot([0.99739 3], idy,'k','LineStyle','-','LineWidth', 4)
hold on;
txt = 'RETENTION';
text(1.1, 0.85*10^-3,txt,'FontSize',35)
hold on
% Add letter B
txt = 'B';
text(0.8,0.5*10^-3,txt,'Fontsize',30)
hold on
return
function [Eqs,Fi] = Call_Fsolve_FiFG_25Mar19(Guess,V1,VarVec)

%-------------------------------------------------------------------
% This is the FUNCTION which returns the Fi for the first gas which
produces a particular value for the gas uptake V1 (STPD).
% Started 20/08/2017
%-------------------------------------------------------------------
% Specify optimization options for Matlab fsolve function (nonlinear

system solver)
% Display options are 'off', 'none', 'final', 'final-detailed',
'iter', or 'iter-detailed')
options1 = optimoptions('fsolve','Algorithm','trust-region-dogleg');
options2 = optimoptions('fsolve','Algorithm','trust-region-
reflective');

options3 = optimoptions('fsolve','Algorithm','Levenberg-Marquardt');
options = optimoptions(options1,'TolFun',1e-16,'TolX',1e-16,...
'Display','none','Maxiter',50,'MaxFunEvals',500);
%options = [];
[Vars] = fsolve(@Get_FiFG_25Mar19,Guess,options,V1,VarVec);
[Eqs,Fi] = Get_FiFG_25Mar19(Vars,V1,VarVec);
end
function varargout = Get_FiFG_25Mar19(Vars,V1,VarVec)

% This function returns the Fi for the FG associated with a given V1
% Written by Ben Korman BSc MBBS MD FANZCA
% Started 08/12/2018
% Fi = Fi1
% V1 = the uptake of FG in L/min
% L = L1, the solubility of FG in blood
% Va = total expired alveolar ventilation in L/min
% Q = total pulmonary blood flow in L/min
% sigma = degree of V/Q mismatch
% Fv = Fv1
% Revised on 23/12/2017 by rewriting the function for Vi to try and
prevent Fi1 exceeding 1
% Last revised 30/10/2018
% A first guess at Fi1 is made in the main program
Va = VarVec(1);
Q = VarVec(2);
sigma = VarVec(3);
L = VarVec(4);
Fv = VarVec(5);
Fi = Vars(1);
fun = @(x) 1./(((1/Q)*exp(((x+sigma/2).^2)/2))+(L*(1-

Fi)*(1/Va)*exp(((x-sigma/2).^2)/2)));
I = integral(fun,-Inf,Inf);
V1_est = L*(Fi-Fv)*I/sqrt(2*pi);
% % Define steady-state gas equation to be solved for unknown

variables
Eqs(1) = V1 - V1_est;
if (nargout == 1)
varargout = {Eqs};
else
varargout = {Eqs,Fi};
end
end

APPENDIX C.3 ROUTINES USED TO DRAW FIGURE 7.7
% This program uses the 2-gas model to determine the compartmental
uptake of a gas present in very low concentrations in the inspired
gas mixture for different degrees of V/Q mismatch.
%
% We graph variables against (1/sigma)log(v/q). This facilitates
% recognition of the site of gas uptake as the maximum of Va and Q
occur at exactly sigma/2 units on opposite sides of the y-axis.
% Written by Dr Ben Korman BSc, MBBS, MD, FFARACS, FANZCA
% Started 8Dec18
% Last Revised 15Sep19

format
V0 = 4; % Expired alveolar ventilation, l/min

Q0 = 5; % Pulmonary blood flow, l/min
L = 13; % Solubility of gas under investigation

Fi2 = 0;% Fractional concentration in inspired gas mixture - not
needed here
Fv = 0; % Fractional concentration in gas sample in eqlm with venous
blood
% Different values of sigma for which uptake is to be calculated

mismatch = 0.01:0.5:2.01;
num_mismatch = length(mismatch);
% Coordinates of median
xG_50_V = zeros(1,num_mismatch);
yG_50_V = zeros(1,num_mismatch);
% Set up partition with N+1 equally spaced points on the x-axis

N = 800;
x = -4:8/N:4;
x_vector = zeros(num_mismatch,N+1);
for i = 1:num_mismatch
x_vector(i,:) = x;
end
y_vector = zeros(num_mismatch,N+1);
sigma = mismatch(i);
mu = sigma/2;
fun = @(x) 1./((L*exp(((x-

mu).^2)/2)+((V0/Q0)*exp(((x+mu).^2)/2))));
I = integral(fun,-Inf,Inf);
y = 1./((L*exp(((x-mu).^2)/2)+((V0/Q0)*exp(((x+mu).^2)/2))));
y_vector(i,:) = y/I;
% Get x value for 50% uptake

% Set Guess according to whether L < V0/Q0 or L > V0/Q0

if L < V0/Q0
Guess = -sigma/4;
else
Guess = sigma/4;
end
VarVec = [V0 Q0 mu L Fv];

[Eqs1, x_50, y_50] = Call_Fsolve_SG_xy_50(Guess,VarVec,I);
xG_50_V(i) = x_50;
yG_50_V(i) = y_50;
end
num_fig = 1;
0.3 20 10]);
max_y = 0.5; % Adjust this value as necessary

col = ['b' 'r' 'g' 'm' 'c' 'y'];
kol = col(i);
y = y_vector(i,:);
plot (x,y,kol)
if i == 1
xlim([-4 4])
ylim([0 max_y])
xlabel('(1/\sigma)log(v/q)')
Title = 'Fraction of Total Gas Uptake as a function of

(1/\sigma)log(v/q)';
ylabel('Fraction of Total Gas Uptake, v_{SG}')
title(Title)
hold on
% Plot vertical line at x = 0
idx = 0;
plot([idx idx],y1,'k','LineStyle', '-','Linewidth',1)
hold on
end
% Plot vertical line at median value of gas uptake

idx = xG_50_V(i);
y_ = yG_50_V(i);
plot([idx idx],[0 y_],kol,'LineStyle', '-','Linewidth',1)
hold on
if i<num_mismatch
pause
end
end
return

function [Eqs,x_50,y_50] = Call_Fsolve_G_xy_50(Guess,Varvec,I)
% This function calculates the value of x = (1/sigma)log(v/q) for

which 50% of uptake occurs

system solver)
reflective');
%options = [];
[Vars] = fsolve(@Get_G_xy_50,Guess,options,Varvec,I);
[Eqs,x_50,y_50] = Get_G_xy_50(Vars,Varvec,I);
End
function varargout = Get_G_xy_50(Vars,Varvec,I)

% This function determines the value of x = (1/s)log(v/q) for which
v has its maximal value and the value of v at this value of x
% Commenced 26Mar19
V0 = Varvec(1);
Q0 = Varvec(2);
mu = Varvec(3);
L = Varvec(4);
V0_Q0 = V0/Q0;
X = Vars(1);
% Calculate I_X, the denominator of v for this value of sigma

fun_g = @(x) 1./(L*(exp(((x-
mu).^2)/2))+V0_Q0*(exp((x+mu).^2)/2)));
I_X = integral(fun_g,-Inf,X)
Eqs(1) = I_X – 0.5*I;
Y = 1/(L*(exp(((X-mu)^2)/2))+V0_Q0*(exp(((X+mu)^2)/2)));
Y = Y/I;
if (nargout == 1)
varargout = {Eqs};
else
varargout = {Eqs,X,Y};
end
end

APPENDIX C.4 ROUTINES USED TO DRAW FIGURE 8.2
% THIS PROGRAM DETERMINES THE PROFILE OF SG UPTAKE AND OF THAT
REMAINING IN THE GAS PHASE FOR A SELECTED VALUE OF SIGMA.
% for CONSTANT INFLOW and CONSTANT OUTFLOW

% We set Va = 4, Qt = 5, Fisg = 0.01 and use sigma = 2
% We then add 70% nitrous oxide and map its uptake to investigated
possible concentrating effects on the SG left in the Gas Phase
% Commenced 31 March 2019

% Last revised 4 April 2019
% Last Revised 1 March 2020

format long
V0 = 4;
Q0 = 5;
V0_Q0 = V0/Q0;
L2ml = 1000;
L1 = 0.47; %0.47;
s = 2; % symbol for sigma

Fv = 0;
Fi = 0.7;
fun = @(x) 1./((L1*(1-Fi)*(exp(((x-
s/2).^2)/2)))+(V0_Q0*(exp(((x+s/2).^2)/2))));
dV = integral(fun,-Inf,Inf);
dV = L1*Fi*V0*dV/sqrt(2*pi);
L1 = 0.47;
%% Set the value of sigma being investigated
s = 2;
VarVec = [V0 Q0 s L1 Fv];
% CONSTANT OUTFLOW
Guess = 0.5;
[Eqs,F1i] = Call_Fsolve_FiFG_25Mar19(Guess,dV,VarVec);
L1p = L1*(1-F1i);
F1v = Fv;
F2i = 0.01;
L2 = 0.42;
dx = 0.001;
x = -4+s/2:0.01:4+s/2;
num_points = length(x);
Va = zeros(1,num_points);
Vi = zeros(1,num_points);
Qi = zeros(1,num_points);

V2r = zeros(1,num_points);
V2tu= zeros(1,num_points);
V2xu= zeros(1,num_points);
V1i =zeros(1,num_points);
V1r=zeros(1,num_points);
z = x(i);
% Work on fixed Vi first
Va(i) = (V0/sqrt(2*pi))*exp((-(z-s/2)^2)/2);
V1u(i) = 1/((L1p*exp(((z-s/2)^2)/2))+V0_Q0*exp(((z+s/2)^2)/2));
V1u(i) = (L1*F1i*V0/sqrt(2*pi))*V1u(i); % uptake of FG
Vi(i) = Va(i)+V1u(i);
Qi(i) = (Q0/sqrt(2*pi))*exp((-(z+s/2)^2)/2);
V2i(i) = Vi(i)*F2i;
V2u(i) = L2*F2i*Vi(i)/(L2 + Vi(i)/Qi(i));
V2r(i) = V2i(i) - V2u(i);
% Now calculate Total uptake of SG

V2tu(i) = L2*F2i*Vi(i)/(L2 + Va(i)/Qi(i));
V2xu(i) = V2tu(i)-V2u(i);
end
[M1, posM1] = max(V1u);

[M2, posM2] = max(V2r);
xM1 = x(posM1);
xM2 = x(posM2);
y_ll = -0.025*L2ml;
y_ul = +0.025*L2ml;
num_fig = 1;
0.5 20 10]);
box on
ylim([y_ll y_ul])
xlim([-3 5])
hold on
plot(x,V2i*L2ml,'b','LineStyle','-','LineWidth',4);
hold on
plot(x,-V2u*L2ml,'r','LineWidth',1);
hold on
plot(x,V2r*L2ml,'k','LineWidth',4);
hold on
% Shade scaled contraction
upperY = ones(1,num_points);
upperY = y_ul*upperY;
fill([x fliplr(x)], [upperY fliplr(upperY-(V2i-F2i*Va)*L2ml)], 'y');
% Shade remaining second gas
fill([x fliplr(x)], [V2r fliplr(V2r-V2r)],[0.94 0.99 0.94]);
% Shade absorbed gas
fill([x fliplr(x)], [-V2u fliplr(V2u-V2u)],'r');
%Draw a horizontal line at y = 0
Xl = x(1);

Xr = x(length(x));
hold on;
plot([Xl Xr], [0 0],'k','LineStyle','-')
hold on
%
xlabel('(1/\sigma) log(v/q)')
ylabel('Second gas flows (ml/min)')
%Title = 'First step of second gas uptake as a function of
(1/\sigma) log(v/q)';
%title(Title)
%
%Draw a vertical line at x = maximum of N2O uptake
plot([xM1 xM1], [(y_ul - M1*L2ml) y_ul], 'k', 'LineStyle','-
','LineWidth', 3)
%Draw a vertical line at x = maximum of SG uptake
','LineWidth', 3)
set(gca,'FontSize',20','LineWidth',3);
xticks([-3 -2 -1 0 1 2 3 4 5])
if L2 == 0.0076
print -dtiff -r600 'Fig8A_CO_L.png'
elseif L2 == 0.42
print -dtiff -r600 'Fig8B_CO_L.png'
else
print -dtiff -r600 'Fig8C_CO_L.png'
end
pause
num_fig = num_fig+1; %Fig 2

%figure(num_fig); set(figure(num_fig),'Units','inches','Position',[0
0.25 10 7.5]);
0.5 20 10]);
box on
ylim([y_ll y_ul])
xlim([-3 5])
hold on
plot(x,V2i*L2ml,'g','LineStyle',':','LineWidth',2);
hold on
plot(x,V2r*L2ml,'b','LineStyle','-','LineWidth',4);
hold on
%plot(x,F2i*Va,'g','LineStyle',':');
hold on
plot(x,-V2xu*L2ml,'k','LineWidth',1);
hold on
plot(x,(V2r-V2xu)*L2ml,'k','LineWidth',4);
hold on
fill([x fliplr(x)], [(V2r-V2xu)*L2ml fliplr(V2r-V2r)],[0.94 0.94
0.94]);
fill([x fliplr(x)], [-V2xu*L2ml fliplr(V2xu-V2xu)],'r');
Xl = x(1);

Xr = x(length(x));
hold on;
hold on
%Title = 'Second step of second gas uptake as a function of
%title(Title)
xticks([-3 -2 -1 0 1 2 3 4 5])
','LineWidth', 3)
','LineWidth', 3)
if L2 == 0.0076
print -dtiff -r600 'Fig8A_CO_R.png'
elseif L2 == 0.42
print -dtiff -r600 'Fig8B_CO_R.png'
else
print -dtiff -r600 'Fig8C_CO_R.png'
end
pause
%% CONSTANT INFLOW
%% Note that Vi now equals V0+dV to exactly match the constant
outflow case.
V0 = V0+dV;
% Start by determining Fi for this value of V_fg = dV
Guess = F1i;
[Eqs, F1i] = Call_FSolve_FG_uptake_ci(Guess,V0,Q0,L1,s,dV);
mu = s/2;
eta = V0/(L1*Q0);
z = x(i);
% Work on fixed Vi first
Vi(i) = (V0/sqrt(2*pi))*exp((-(z-s/2)^2)/2);
V1u(i) = exp(-(z+mu)^2/2) + eta*exp(-(z-mu)^2/2);

V1u(i) = V1u(i) - sqrt(V1u(i)^2-4*eta*F1i*exp(-(z^2+mu^2)));
V1u(i) = (L1*Q0/(2*sqrt(2*pi)))*V1u(i); % distribution of uptake
of FG
Va(i) = Vi(i)-V1u(i);

Qi(i) = (Q0/sqrt(2*pi))*exp((-(z+mu)^2)/2);
V2i(i) = Vi(i)*F2i;
V2u(i) = L2*F2i*Vi(i)/(L2 + Vi(i)/Qi(i));
V2r(i) = V2i(i) - V2u(i);

% Now calculate Total uptake of SG
V2tu(i) = L2*F2i*Vi(i)/(L2 + Va(i)/Qi(i));
V2xu(i) = V2tu(i)-V2u(i);
end
[M1, posM1] = max(V1u);

[M2, posM2] = max(V2r);
xM1 = x(posM1);
xM2 = x(posM2);
y_ll = -0.025*L2ml;
y_ul = +0.025*L2ml;
num_fig = num_fig + 1; %Fig 3
0.5 20 10]);
ylim([y_ll y_ul])
hold on
plot(x,V2i*L2ml,'b','LineStyle','-','LineWidth',4);
hold on
plot(x,-V2u*L2ml,'k','LineWidth',1);
hold on
plot(x,V2r*L2ml,'k','LineWidth',4);
hold on
fill([x fliplr(x)], [V2r*L2ml fliplr(V2r-V2r)],[0.94 0.99 0.94]);
fill([x fliplr(x)], [-V2u*L2ml fliplr(V2u-V2u)],'r');
Xl = x(1);
Xr = x(length(x));
hold on;
hold on
%
ylabel('Second gas flows')
%Title = 'First step of second gas uptake as a function of
%title(Title)
hold on
if L2 == 0.0076
print -dtiff -r600 'Fig8A_CI_L.png'
elseif L2 == 0.42
print -dtiff -r600 'Fig8B_CI_L.png'
else
print -dtiff -r600 'Fig8C_CI_L.png'
end
pause
num_fig = num_fig+1; %Fig 4

%figure(num_fig); set(figure(num_fig),'Units','inches','Position',[0
0.25 10 7.5]);
0.5 20 10]);
box on

ylim([y_ll y_ul])
hold on
plot(x,V2i*L2ml,'g','LineStyle',':','LineWidth',2);
plot(x,V2r*L2ml,'b','LineStyle','-','LineWidth',4);
hold on
%plot(x,F2i*Va,'g','LineStyle',':');
hold on
plot(x,-V2xu*L2ml,'k','LineWidth',1);
hold on
plot(x,(V2r-V2xu)*L2ml,'k','LineWidth',4);
hold on
%fill([x fliplr(x)], [V2i fliplr(F2i*Va)], 'y');
fill([x fliplr(x)], [(V2r-V2xu)*L2ml fliplr(V2r-V2r)],[0.94 0.94
0.94]);
fill([x fliplr(x)], [-V2xu*L2ml fliplr(V2xu-V2xu)],'r');
Xl = x(1);
Xr = x(length(x));
hold on;
hold on
Title = 'Second step of second gas uptake as a function of
%title(Title)
xticks([-3 -2 -1 0 1 2 3 4 5])
','LineWidth', 3)
','LineWidth', 3)
if L2 == 0.0076
print -dtiff -r600 'Fig8A_CI_R.png'
elseif L2 == 0.42
print -dtiff -r600 'Fig8B_CI_R.png'
else
print -dtiff -r600 'Fig8C_CI_R.png'
end
return

function [Eqs,Fi] = Call_Fsolve_FiFG_25Mar19(Guess,V1,VarVec)
%------------------------------------------------------------------
% This is the FUNCTION which returns the Fi for the first gas which
produces a particular value for the uptake of first gas V1 (STPD).
% Started 20/08/2017
%------------------------------------------------------------------

system solver)
reflective');
%options = [];
[Vars] = fsolve(@Get_FiFG_25Mar19,Guess,options,V1,VarVec);
[Eqs,Fi] = Get_FiFG_25Mar19(Vars,V1,VarVec);
end

Bibliography
1. Anderson JC, Babb AL, and Hlastala MP. Modeling soluble gas exchange
in the airways and alveoli. Ann Biomed Eng 31: 1402-1422, 2003.
2. Anderson JC, and Hlastala MP. Breath tests and airway gas exchange. Pulm
Pharmacol Ther 20: 112-117, 2007.
3. Aranake A, Mashour GA, and Avidan MS. Minimum alveolar
concentration: ongoing relevance and clinical utility. Anaesthesia 68: 512-522, 2013.
4. Bartle RG. A Modern Theory of Integration. American Mathematical Society,
2001, p. 458.
5. Biro GP. From the Atmosphere to the Mitochondrion: The Oxygen Cascade.
In: Hemoglobin-Based Oxygen Carriers as Red Cell Substitutes and Oxygen
Therapeutics, edited by Kim H, and Greenburg A. Berlin Heidelberg: Springer, 2013,
p. 27-53.
6. Bloom WR, and Boyce P. Alexander Proban Robertson. Aust Math Soc
Gazette 22: Aust MS: Homepage, 1995.
7. Calvey TN, and Williams NE. Inhalational Anaesthetic Agents. In: Principle
and Practice of Pharmacology for Anaesthetists. Massachusetts: Blackwell
Publishing, Inc., 2008, p. 136-137.
8. Canfield RE, and Rahn H. Arterial-alveolar N2 gas pressure differences due
to ventilation-perfusion differences. J Appl Physiol 10: 165-172, 1957.
9. Chinard FP, Enns T, and Nolan MF. Diffusion and solubility factors in
pulmonary inert gas exchanges. J Appl Physiol 16: 831-836, 1961.
Bibliography 205
10. Colburn WE, Evans JW, and West JB. Analysis of effect of the solubility
on gas exchange in nonhomogenous lungs. J Appl Physiol 37: 547-551, 1974.
11. Community EE. Options to reduce nitrous oxide emissions (Final Report)
European Econmic Community, 1998.
12. Dantzker DR, Wagner PD, and West JB. Instability of lung units with low
VA/Q ratios during O2 breathing. Journal of applied physiology 38: 886-895, 1975.
13. Dash RK, and Bassingthwaighte JB. Erratum to: Blood HbO2 and HbCO2
dissociation curves at varied O2, CO2, pH, 2,3-DPG and temperature levels. Ann
Biomed Eng 38: 1683-1701, 2010.
14. Dash RK, Korman B, and Bassingthwaighte JB. Simple accurate
mathematical models of blood HbO2 and HbCO2 dissociation curves at varied
physiological conditions – evaluation and comparison with other models. Eur J Appl
Physiol 116: 97-113, 2016.
15. Divatia JV, Vaidya JS, Badwe RA, and Hawaldar RW. Omission of nitrous
oxide during anesthesia reduces the incidence of postoperative nausea and vomiting.
Anesthesiology 85: 1055-1062, 1996.
16. Dueck R, Young I, Clausen J, and Wagner PD. Altered distribution of
pulmonary ventilation and blood flow following induction of inhalational anesthesia.
17. Ebert TJ, and Naze SA. Inhaled Anesthetics. In: Clinical Anesthesia, edited
by Barash PG, Calahan MK, Cullen BF, Holt NF, Ortega R, Sharar SR, Stock MC,
and Stoelting RK. Philadelphia: Wolters Kluwer, 2017, p. 459-483.
18. Eger EI. Applications of a mathematical model of gas uptake. In: Uptake and
Distribution of Anesthetic Agents, edited by Papper EM, and Kitz RJ. New York:
McGraw-Hill, 1963, p. 88-103.
206 Bibliography
19. Eger EI. Effect of inspired anesthetic concentrations on rate of rise of alveolar
concentration. Anesthesiology 24: 153-157, 1963.
20. Eger EI. Inhaled anesthetics:uptake and distribution. In: Miller's Anesthesia,
edited by Miller R. Philadelphia: Elsevier Health Sciences, 2010, p. 543.
21. Eger EI. A mathematical model of uptake and distribution. In: Uptake and
Distribution of Anesthetic Agents, edited by Papper EM, and Kitz RJ. New York:
McGraw-Hill, 1963, p. 72-87.
22. Eger EI. Uptake, distribution, and elimination of inhaled anaesthetics. In:
Scientific Foundations of Anaesthesia, edited by Scurr C, and Feldman S. London:
William Heinemann, 1974, p. 444-453.
23. Eger EI, Tang M, Liao M, Laster MJ, Solt K, Flood P, Jenkins A, Raines
D, Hendrickx JFA, Shafer SL, Yasumasa T, and Sonner JM. Inhaled anesthetics
do not combine to produce synergistic effects regarding minimum alveolar anesthetic
concentration in rats. Anesth Analg 107: 479-485, 2008.
24. Epstein RM, Rackow H, Salanitre E, and Wolf GL. Influence of the
Concentration Effect on the Uptake of Anesthetic Mixtures The Second Gas Effect.
25. Evans JW. Mathematical analysis of compartmental lung models. In:
Pulmonary Gas Exchange, edited by West JB. New York, New York: Academic Press,
Inc., 1980, p. 308-334.
26. Evans JW, Wagner PD, and West JB. Conditions for reduction of pulmonary
gas transfer by ventilation-perfusion inequality. J Appl Physiol 36: 533-537, 1974.
27. Farhi LE, and Olszowka AJ. Analysis of alveolar gas exchange in the
presence of soluble inert gases. Respir Physiol 5: 53-67, 1968.
Bibliography 207
28. Flippo TS, and Holder Jr WD. Neurologic degeneration associated with
nitrous oxide anesthesia in patients with vitamin B12 deficiency. Archives of Surgery
128: 1391-1395, 1993.
29. Forman SA, and Benkwitz C. Inhalational Anesthetics. In: Anesthesiology,
edited by Longnecker DE, Mackey SC, Newman MF, Sandberg WS, and Zapol WM.
China: McGraw-Hill Education, 2018, p. 551-570.
30. Forster RE. Diffusion factors in gases and liquids. In: Uptake and Distribution
of Anesthetic Agents, edited by Papper EM, and Kitz RJ. New York: McGraw-Hill
Book Co. Inc., 1963, p. 20-29.
31. Frumin MJ, Salanitre E, and Rackow H. Excretion of nitrous oxide in
anesthetized man. J Appl Physiol 16: 720-722, 1961.
32. Hartung J. Twenty-four of twenty-seven studies show a greater incidence of
emesis associated with nitrous oxide than with alternative anesthetics. Anesthesia &
Analgesia 83: 114-116, 1996.
33. Heller ML, Watson TRJ, and Imredy DS. Effect of nitrous oxide uptake on
arterial oxygenation. Anesthesiology 28: 904-913, 1967.
34. Hendrickx JF, Carette R, Lemmens HJM, and De Wolf AM. Large volume
N2O uptake alone does not explain the second gas effect of N2O on sevoflurane during
constant inspired ventilation. Br J Anaesth 96: 391-395, 2006.
35. Hunter JM, and Hemmings HC. The many contributions of WW Mapleson
to the British Journal of Anaesthesia. Br J Anaesth 122: 159-162, 2019.
36. Karris S. Numerical Analysis Using MATLAB and Excel. USA: Orchard
Publications, 2007, p. 391-393.
37. Katoh T, Ikeda K, and Bito H. Does nitrous oxide antagonize sevoflurane-
induced hypnosis? Br J Anaesth 79: 1997.
208 Bibliography
38. Kelman GR. Digital computer procedure for the conversion of PCO2 into blood
CO2 content. Respir Physiol 3: 111-115, 1967.
39. Kelman GR. Digital computer subroutine for the conversion of oxygen tension
into saturation. J Appl Physiol 21: 1375-1376, 1966.
40. Kelman GR. A new lung model: an investigation with the aid of a digital
computer. Comp Biomed Res 3: 241-248, 1970.
41. Kennedy RR. A Second Look at the Second Gas Effect. Anesthesiology 128:
1053-1054, 2018.
42. Kety SS. The theory and applications of the exchange of inert gas at the lungs
and tissues. Pharmacol Rev 3: 1-41, 1951.
43. Kitihata LM, Taub A, and Conte AJ. The effect of nitrous oxide on alveolar
carbon dioxide tension: a second-gas effect. Anesthesiology 35: 607-611, 1971.
44. Korman B. Binary Solutions in Anaesthesia. In: MD thesis, Melbourne, 1994,
p. 166.
45. Korman B, Dash RK, and Peyton PJ. Can mathematical modeling explain
the measured magnitude of the second gas effect? Anesthesiology 128: 1075 - 1083,
2018.
46. Korman B, Dash RK, and Peyton PJ. Effect of net gas volume changes on
alveolar and arterial gas partial pressures in the presence of ventilation-perfusion
mismatch. J Appl Physiol 126: 558-568, 2019.
47. Korman B, Dash RK, and Peyton PJ. Elucidating the roles of solubility and
ventilation-perfusion mismatch in the second gas effect using a two-step model of gas
exchange. J Appl Physiol 128: 1587-1593, 2020.
48. Korman B, and Mapleson WW. Concentration and second gas effects: can
the accepted explanation be improved? Br J Anaesth 78: 618-625, 1997.
Bibliography 209
49. Kretzschmar M, Kozian A, Baumgardner JE, Batista Borges J, and
Hedenstierna G. The Effect of Bronchoconstriction-induced Ventilation/Perfusion-
Mismatch on Uptake and Elimination of Isoflurane and Desflurane. Anesthesiology
127: 800-812, 2017.
50. Lee ASJ, and Sun XG. Response to Letter to the Editor. Anesthesia &
Analgesia 89: 1326, 1999.
51. Leslie K, Myles PS, Chan MTV, Forbes A, Paech MJ, Peyton PJ, and
Silbert BS. Nitrous oxide and long-term morbidity and mortality in the ENIGMA trial.
Anesth Analg 112: 387-393, 2011.
52. Lundh R, and Hedenstierna G. Ventilation-perfusion relationships during
halothane anesthesia and mechanical ventilation. Effects of varying inspired oxygen
concentration. Acta Anaesthesiol Scand 28: 191-198, 1984.
53. Lundh R, and Hedenstierna G. Ventilation‐perfusion relationships during
anaesthesia and abdominal surgery. Acta Anaesthesiol Scand 27: 167-173, 1983.
54. Lundh R, and Johansson H. Alveolar stability during anaesthesia for
reconstructive vascular surgery in the leg. Acta Anaesthesiol Scand 27: 26-34, 1983.
55. Magnusson L, and Spahn DR. New concepts of atelectasis during general
anaesthesia. Br J Anaesth 91: 61-72, 2003.
56. Mapleson WW. Circulation-time models of the uptake of inhaled anaesthetics
and data for quantifying them. Br J Anaesth 45: 319-334, 1973.
57. Mapleson WW. An electric analogue for uptake and exchange of inert gases
and other agents. J Appl Physiol 18: 197-204, 1963.
58. Mapleson WW. Inert gas-exchange theory using an electric analogue. J Appl
Physiol 19: 1193-1198, 1964.
210 Bibliography
59. Mapleson WW, and Korman B. The Second Gas Effect Is a Valid Concept.
Anesth Analg 89: 1326, 1999.
60. McShane EJ. Integration. Princeton, N.J.: Princeton Univ Press, 1944.
61. Merkel G, and Eger EI. A comparative study of halothane and halopropane
anesthesia. Anesthesiology 24: 346-357, 1963.
62. Myles PS, Leslie K, Chan MTV, Forbes A, Paech MJ, Peyton PJ, Silbert
BS, and Pascoe E. Avoidance of nitrous oxide for patients undergoing major surgery:
a randomized controlled trial. Anesthesiology 107: 221-231, 2007.
63. Myles PS, Leslie K, Chan MTV, Forbes A, Peyton PJ, Paech MJ, Beattie
WS, Sessler DI, Devereaux PJ, and Silbert BS. The safety of addition of nitrous
oxide to general anaesthesia in at-risk patients having major non-cardiac surgery
(ENIGMA-II): a randomised, single-blind trial. The Lancet 384: 1446-1454, 2014.
64. Neufeld GR, Williams JJ, Klineberg PL, and Marshall BE. Inert gas a-A
differences: a direct reflection of V/Q distribution. J Appl Physiol 44: 277-283, 1978.
65. Nishikawa K, Kunimoto F, Isa Y, Miyoshi S, Takahashi K, Morita T, Arii
H, and Goto F. Second gas effect of N2O on oxygen uptake. Can J Anesth 47: 506-
510, 2000.
66. Peyton P. Physiological determinants during anaesthesia of the accuracy of
pulmonary blood flow measurement from soluble gas exchange. In: MD thesis,
Melbourne, 2004.
67. Peyton P, Horriat M, Robinson G, Pierce R, and Thompson B. Magnitude
of the second gas effect on arterial sevoflurane partial pressure. Anesthesiology 108:
381-387, 2008.
Bibliography 211
68. Peyton P, Robinson G, and Thompson B. Effect of ventilation-perfusion
inhomogeneity and N2O on oxygenation: physiological modeling of gas exchange. J
Appl Physiol 91: 17-25, 2001.
69. Peyton P, Robinson G, and Thompson B. Ventilation-perfusion
inhomogeneity increases gas uptake in anesthesia: computer modeling of gas
exchange. J Appl Physiol 91: 10-16, 2001.
70. Peyton P, Robinson G, and Thompson B. Ventilation-perfusion
inhomogeneity increases gas uptake: theoretical modeling of gas exchange. J Appl
Physiol 91: 3-9, 2001.
71. Peyton P, Stuart‐Andrews C, Deo K, Strahan F, Robinson G, Thompson
B, and Pierce R. Persisting concentrating and second gas effects on oxygenation
during N2O anaesthesia. Anaesthesia 61: 322-329, 2006.
72. Peyton PJ, Fortuin M, Robinson GJB, Stuart‐Andrews C, Pierce R, and
Thompson BR. The rate of alveolar‐capillary uptake of sevoflurane and nitrous oxide
following anaesthetic induction. Anaesthesia 63: 358-363, 2008.
73. Power GP, Ritchie MA, Drok KJ, and Macleod ID. Ian Mackay Ritchie
1936-2014. Historical Records of Australian Science 28: 50-57, 2017.
74. Rackow H. Discussion of Lung Factors, Applications of a Mathematical
Model of Gas Uptake. In: Uptake and Distribution of Anesthetic Agents, edited by
Papper EM, and Kitz RJ. New York: McGraw-Hill Book Company, Inc., 1963, p. 98-
103.
75. Rackow H, Salanitre E, and Frumin MJ. Dilution of alveolar gases during
nitrous oxide excretion in man. J Appl Physiol 16: 723-728, 1961.
76. Rahn H, and Fenn WO. A Graphical Analysis of the Respiratory Gas
Exchange. Washington, D.C.: The American Physiological Society, 1955, p. 39.
212 Bibliography
77. Ravishankara AR, Daniel JS, and Portmann RW. Nitrous oxide (N2O): the
dominant ozone-depleting substance emitted in the 21st century. Science 326: 123-
125, 2009.
78. Rehder K, Knopp TJ, Sessler AD, and Didier EP. Ventilation-perfusion
relationship in young healthy awake and anesthetized-paralyzed man. J Appl Physiol
Respir Environ Exerc Physiol 47: 745-753, 1979.
79. Riley RL, and Cournand A. "Idea" alveolar air and the analysis of ventilation-
perfusion relationships in the lungs. J Appl Physiol 1: 825-847, 1949.
80. Robinson GJB, Peyton PJ, Vartuli GM, Burfoot RB, and Junor PA.
Continuous measurement of cardiac output by inert gas throughflow: comparison with
thermodilution. Journal of Cardiothoracic and Vascular Anesthesia 17: 204-210,
2003.
81. Salanitre E, and Rackow H. Recent advances in uptake and excretion of
inhalation anesthetics. In: Clinical Anesthesia - A Decade of Clinical Progress, edited
by Fabian LW. Philadelphia: F.A. Davis, 1971, p. 179-185.
82. Severinghaus JW. Can large volume N2O uptake alone explain the second gas
effect? Brit J Anaesth 97: 262, 2006.
83. Severinghaus JW. The rate of uptake of nitrous oxide in man. J Clin Invest
33: 1183-1189, 1954.
84. Stoelting RK, and Eger EI. An Additional Explanation for the Second Gas
Effect: A Concentrating Effect. Anesthesiology 30: 273-277, 1969.
85. Sun XG, Su F, Shi YQ, and Lee C. The" second gas effect" is not a valid
concept. Anesth Analg 88: 188-192, 1999.
Bibliography 213
86. Taheri S, and Eger EI. A demonstration of the concentration and second gas
effects in humans anesthetized with nitrous oxide and desflurane. Anesth Analg 89:
774-780, 1999.
87. Tramer M, Moore A, and McQuay H. Omitting nitrous oxide in general
anaesthesia: meta-analysis of intraoperative awareness and postoperative emesis in
randomized controlled trials. British Journal of Anaesthesia 76: 186-193, 1996.
88. Wagner PD, Saltzman HA, and West JB. Measurement of continuous
distributions of ventilation-perfusion ratios: theory. J Appl Physiol 36: 588-599, 1974.
89. West J, Wagner P, and Derks C. Gas exchange in distributions of Va/Q
ratios: partial pressure-solubility diagram. J Appl Physiol 37: 533-540, 1974.
90. West JB. Effect of slope and shape of dissociation curve on pulmonary gas
exchange. Respir Physiol 8: 66-85, 1969.
91. West JB. Ventilation-perfusion inequality and overall gas exchange in
computer models of the lung. Respiration Physiology 7: 88-110, 1969.
92. West JB, and Wagner PD. Pulmonary gas exchange. In: Bioengineering
aspects of the lung, edited by West JB. New York: Marcel Dekker Inc, 1977, p. 378-
382.
214 Bibliography
Can Mathematical Modeling Explain the Measured
Magnitude of the Second Gas Effect?
Ben Korman, M.D., Ranjan K. Dash, Ph.D., Philip J. Peyton, M.D.
ABSTRACT
Background: Recent clinical studies suggest that the magnitude of the second gas effect is considerably greater on arterial
blood partial pressures of volatile agents than on end-expired partial pressures, and a significant second gas effect on blood
partial pressures of oxygen and volatile agents occurs even at relatively low rates of nitrous oxide uptake. We set out to further
investigate the mechanism of this phenomenon with the help of mathematical modeling.
Methods: Log-normal distributions of ventilation and blood flow were generated representing the range of ventilation-perfusion
scatter seen in patients during general anesthesia. Mixtures of nominal delivered concentrations of volatile agents (desflurane,
isoflurane and diethyl ether) with and without 70% nitrous oxide were mathematically modeled using steady state mass-balance
principles, and the magnitude of the second gas effect calculated as an augmentation ratio for the volatile agent, defined as the
partial pressure in the presence to that in the absence of nitrous oxide.
Results: Increasing the degree of mismatch increased the second gas effect in blood. Simultaneously, the second gas effect
decreased in the gas phase. The increase in blood was greatest for the least soluble gas, desflurane, and least for the most soluble
gas, diethyl ether, while opposite results applied in the gas phase.
Conclusions: Modeling of ventilation-perfusion inhomogeneity confirms that the second gas effect is greater in blood than in
expired gas. Gas-based minimum alveolar concentration readings may therefore underestimate the depth of anesthesia during
nitrous oxide anesthesia with volatile agents. The effect on minimum alveolar concentration is likely to be most pronounced
for the less soluble volatile agents in current use. (ANESTHESIOLOGY 2018; 128:1075-83)
T HE second gas effect occurs when a soluble first gas

such as nitrous oxide is delivered in high inspired
concentrations. Alveolar-capillary uptake of the first gas
What We Already Know about This Topic
• The second gas effect occurs when a soluble first gas such
increases the alveolar concentrations of other gases present, as nitrous oxide is delivered in high inspired concentrations
and the alveolar-capillary uptake of the first gas increases the
accelerating their uptake. The effect was first described by alveolar concentrations of other gases present, accelerating
Epstein et al.,1 who believed it was due to extra gas drawn their uptake
into the lung to replace the nitrous oxide taken up by blood. • The magnitude of the second gas effect may be greater when
Subsequently, Stoelting and Eger showed that the second the effect on arterial blood partial pressures is measured due
to the effect of inhomogeneity of ventilation and blood flow
gas effect must involve a concentrating effect as well, i.e., ratios on the distribution of perfusion-driven nitrous oxide
an effect due to volume “shrinkage.”1,2 They produced a uptake throughout the lung
diagram to explain these two steps, a diagram that then
What This Article Tells Us That Is New
appeared in many anesthetic textbooks. Years later, Kor-
man and Mapleson showed that the effect is primarily one • Modeling of ventilation-perfusion inhomogeneity confirmed
that the second gas effect is greater in blood than it is in expired
of volume “shrinkage” and produced modifications of the gas, and its magnitude increases in blood but decreases in
Stoelting-Eger diagram that explained the phenomenon expired gas as the degree of ventilation-perfusion mismatch
more accurately, including equilibration of the alveolar increases
compartment with blood.3 • Minimum alveolar concentration calculations based on
end-tidal anesthetic concentration measurements may well
Several clinical studies involving the measurement of underestimate the depth of anesthesia when nitrous oxide is
end-expired concentrations of volatile anesthetic agents supplemented with a volatile agent
in patients have confirmed the existence of the second gas
effect.4,5 However, in a clinical study prompted by results of the distribution of perfusion-driven nitrous oxide uptake
computer modeling of gas-exchange in the lung,6–8 Peyton throughout the lung, with more powerful concentrating
et al. found that the magnitude of the second gas effect is effects generated in lung compartments with higher blood
considerably greater where the effect on arterial blood par- flow and low V! / Q! . They suggested furthermore that this
tial pressures is measured instead.9 They explained this by could produce a significant second gas effect on blood par-
the effect of inhomogeneity or scatter of ventilation and tial pressures of oxygen and volatile agent even at relatively
blood flow ratios (ventilation-perfusion [V! / Q! ] ratios) on low “maintenance phase” rates of nitrous oxide uptake,10 a
This article is featured in “This Month in Anesthesiology,” page 1A. Corresponding article on page 1053.
Submitted for publication October 21, 2017. Accepted for publication January 12, 2018. From the Department of Anaesthesia and Pain
Medicine, Royal Perth Hospital, Perth, Australia (B.K.); the Departments of Biomedical Engineering and Physiology, Medical College of
Wisconsin, Milwaukee, Wisconsin (R.K.D.); the Anaesthesia, Perioperative and Pain Medicine Unit, Melbourne Medical School, University of
Melbourne, Parkville, Australia (P.J.P.); and the Department of Anaesthesia, Austin Health, Heidelberg, Australia (P.J.P.).
Copyright © 2018, the American Society of Anesthesiologists, Inc. Wolters Kluwer Health, Inc. All Rights Reserved. Anesthesiology 2018; 128:1075-83
Anesthesiology, V 128 • No 6 1075 June 2018
Copyright © 2018, the American Society of Anesthesiologists, Inc. Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
<zdoi;10.1097/ALN.0000000000002131>
Ventilation-Perfusion Mismatch and the Second Gas Effect
suggestion that has been questioned,11 but is supported by constant. With no second gas effect, the augmentation ratio is
clinical studies.5,9 1. Once the second gas effect becomes evident, the augmenta-
The idea that a phenomenon traditionally believed to be tion ratio exceeds 1. The term “augmentation ratio” was first
caused by the large reductions in gas volume associated with used by Epstein et al.1 Although they only calculated augmen-
high rates of nitrous oxide uptake can persist at clinically sig- tation ratios in end-tidal gas, the ratio may also be calculated
nificant concentrations even when those volume changes fall for blood.
away is a significant departure from traditional teaching. We We have assumed a log-normal distribution of V! / Q!
set out to further explore the mechanism of this proposed phe- ratios in our investigations.14 Colburn et al.15 have shown
nomenon with computer modeling of theoretical distributions that if σV! is the log SD of ventilation per unit volume and
of V! / Q! scatter seen in patients during general anesthesia, σ Q! is the log SD of blood flow per unit volume, then σ,
across a range of nitrous oxide uptake rates, and the predicted the degree of mismatch between ventilation and blood flow,
effect on second gases with different solubilities in blood. depends only on σV! − σ Q! , the absolute value of the differ-
ence between the two standard deviations. This is illustrated
Materials and Methods in figure 1. In figure 1A, the ventilation as a fraction of the
The minimum number of components in the gas phase neces- total alveolar ventilation (!v) has been drawn in blue while the
sary for a theoretical study of the second gas effect is as follows: blood flow as a fraction of the total pulmonary blood flow
! has been drawn in red. The unit on the x-axis is not log
(q)
A first gas. Alveolar-capillary uptake of this gas generates the (V! / Q! ) but (1/σ)log( v! / q! ), where σ = σV! − σ Q! . Drawn in
volume changes responsible for the second gas effect.
It must be soluble in blood. The difference in partial this way, the graphs of ventilation and blood flow are exactly
pressures between the inspired gas mixture and mixed the same size and shape, and do not change size or shape as
venous blood must be sufficient to produce the desired σ is varied. What does change is the distance between the
volume changes. mean for ventilation and the mean for blood flow. These
A second gas. This is usually a volatile agent used in low con- means are situated exactly σ/2 units on opposite sides of the
centrations. This gas responds to the volume changes by y-axis. When σ = 0, the curves overlap completely so that
exhibiting the second gas effect. matching between ventilation and blood flow is optimal. As
The remaining gas. This acts purely as a “vehicle” for deliver- σ increases, they overlap less and less and the degree of V! / Q!
ing the first and second gases, and for the purposes of mismatch increases. This is shown further in figure 1B, in
the analysis is regarded as being completely insoluble which the blood flow as a fraction of the total pulmonary
in blood. blood flow (q)! has been rotated through 180 degrees about
the x-axis. In awake healthy adults, σ varies from 0.25 to
In clinical practice, the first gas is always nitrous oxide (Ost- 1.0. Worsening of V! / Q! matching is known to occur soon
wald solubility coefficient in blood at 37°C, λ = 0.47). A after induction of anesthesia and has been well documented
fractional concentration of 70% in dry gas is assumed in all experimentally, with typical values for σ of 0.75 to 1.5.16–20
calculations. This is usually regarded as the highest clinically Further information regarding this way of presenting a log-
effective concentration that can be used safely in the general normal distribution of ventilation and blood flow is pro-
population during anesthesia without producing hypoxia.12 vided in appendix 1.
In modern clinical anesthesia, the second gases used are A general equation for the augmentation ratio (AR) was
mainly desflurane, sevoflurane, and isoflurane. We will con- obtained having the following form in which the difference
sider diethyl ether as well because it is an example of a vola- between inspired alveolar ventilation, V!I , and expired alveo-
tile anesthetic with a high solubility in blood. We examine the lar ventilation, V!A, appears:
second gas effect as the uptake of the first gas, nitrous oxide,
decreases toward zero, as is seen with time in clinical use. This is (V!I − V!A )
achieved in our model by increasing the mixed venous pressure AR = 1 + k (1)
V!
A
of nitrous oxide in a stepwise fashion. The proportion of nitrous
oxide washin is expressed as a washin ratio, which mirrors the Derivation of equation 1 is given in appendix 2. In the absence
usual alveolar/inspired concentration ratio, i.e., a washin ratio of nitrous oxide, inspired alveolar ventilation and expired
of 0.9 is roughly equivalent to the situation where the alveolar/ alveolar ventilation are equal, the augmentation ratio equals
inspired concentration ratio equals 0.9. We estimate that our 1, and there is no augmentation. In the presence of nitrous
steady state equations are reasonably accurate within 10 to 15 oxide, inspired alveolar ventilation is greater than expired
min from the start of induction, when nitrous oxide washin is alveolar ventilation, the augmentation ratio is greater than
more than 90% complete and its uptake rate has plateaued.13 1, and the second gas effect appears. Its magnitude is then
The second gas effect is expressed as an augmentation ratio, (
given by the augmentation ratio. The difference V!I − V!A is )
defined as the partial pressure of second gas in the presence
of nitrous oxide, divided by that which would have existed in
(
equal to the rate of nitrous oxide uptake. The V!I − V!A / V!A )
the absence of the nitrous oxide, with all other variables kept term is the “concentrating effect” or “shrinkage” factor given
Anesthesiology 2018; 128:1075-83 1076 Korman et al.
PERIOPERATIVE MEDICINE
distribution of ventilation and perfusion with σ varying from

0.25 to 2 in increments of 0.25 units. Calculations were per-
formed in both the presence and absence of nitrous oxide.
Outputs were calculated from the flow-weighted uptake rates
of each compartment. The output variable for each second
gas was its augmentation ratio in the blood and gas phases. In
each situation, results were obtained for 3 gases representing a
wide range of blood solubilities: desflurane (λ = 0.42), isoflu-
rane (λ = 1.4), and diethyl ether (λ = 12.1).
Results
Figure 2 shows the results for the blood and gas phases. The
arrow in each diagram of figure 2 indicates the direction in
which σ is increasing. The x-axis shows the proportion of
nitrous oxide washin, and the y-axis shows the augmentation
ratio. The most consistent feature of these diagrams is the
divergence of the lines from the point (1, 1), the point on
the x-axis representing complete washin of the nitrous oxide,
at which time expired alveolar ventilation becomes equal to
inspired alveolar ventilation, so that according to equation 1,
Fig. 1. Log-normal distribution of ventilation and blood flow the second gas effect disappears completely.
drawn so as to show the effect of σ on the degree of match- Each augmentation ratio plot appears to be a straight line.
ing between the two. σ is defined as the absolute value of the The line for σ = 0, which represents a homogenous lung, is
difference between the log SDs of ventilation and blood flow
the same for both gas phase and blood phase, as expected for
each expressed per unit volume (see discussion in Materi-
als and Methods). (A) The y-axis shows the fractional ven-
a “single compartment” lung with full equilibration between
tilation V! / V!A and fractional blood flow Q! / Q! T (represented blood and gas phases. In addition, it is the same for all three
by the symbols v! and q! , respectively); the x-axis is in units gases, i.e., it is independent of the solubility of the second
of (1/σ)log( v! / q! ). The fractional ventilation is shown in blue, gas in blood.
the fractional blood flow in red. For the purpose of this il- The effect of increasing σ and hence the degree of V! / Q!
lustration, σ has been set equal to 2. The figure shows how mismatch is seen to be opposite for the blood and gas phases.
the means of v! and q! are spaced σ/2 units on opposite
In blood, an increase in σ increases the augmentation ratio,
sides of the y-axis (green dashed lines). (B) Here the curve
for blood flow has been rotated by 180 degrees around while in the gas phase, the augmentation ratio decreases.
the x-axis so that −q! is plotted against (1/σ)log( v! / q! ). Moreover, in blood, this amplification is greatest for the least
The region available for gas exchange is located between soluble gas, desflurane, and least for the most soluble gas,
these two graphs. The vertical strip bordered by a dashed diethyl ether, while in the gas phase, the opposite is true.
line represents a single lung compartment in which ventilation
(blue stippled area above the x-axis) has been equilibrated
with blood flow (red stippled area below the x-axis). With Discussion
σ = 0, the graphs of v! and −q! are mirror images, and match- Our mathematical model allows contributions from solu-
ing between ventilation and blood flow is optimal. bility and V! / Q! mismatch to the second gas effect to be
assessed separately. The results show that an increase in
in classical descriptions of the second gas effect. The k term is V! / Q! mismatch augments the second gas effect in blood,
positive and includes contributions from the solubility of the but reduces the second gas effect in the gas phase. This result
is most pronounced for the insoluble gases used in clinical
second gas, the overall alveolar ventilation-perfusion ratio of
practice today. Although not shown here for brevity, results
the lung, and the degree of mismatch between ventilation
obtainable for sevoflurane ( λ = 0.67) are similar to those
and blood flow.
shown for desflurane.
Input variables for each gas were its inspired concentration, The choice of a log-normal distribution is mathemati-
mixed venous partial pressure, and solubility in blood. For sim- cally convenient and has been used previously in simulations
plicity, we have assumed trace concentrations of second gas in of gas exchange in the lungs.6–8,14,21 However, the method
the inspired gas mixture with mixed venous concentrations of used to display the distributions of ventilation and blood
zero. An expired alveolar ventilation of 4 l/min was assumed. flow in figure 1 is new, and illustrates clearly how σ controls
The pulmonary blood flow was set at 5 l/min. Equation 1 was the degree of V! / Q! mismatch. A further advantage of this
first solved for each second gas in a homogenous lung (i.e., method is that it facilitates equilibration of gas and blood
σ = 0), then in a multicompartment lung for a log-normal over the whole range of V! / Q! values, particularly at the
Fig. 2. Second gas effect for desflurane, isoflurane, and diethyl ether as a function of the proportion of washin of nitrous oxide
for different values of σ. The second gas effect is expressed as an augmentation ratio (AR), defined as the partial pressure of the
second gas in the presence of nitrous oxide (N2O), divided by that which would have existed in the absence of the nitrous oxide,
with all other factors kept constant. Upper three diagrams, second gas effect in the gas phase; lower three diagrams, second gas
effect in blood. In each case, the augmentation ratio is plotted on the y-axis as a function of the washin ratio. The long arrow indi-
cates the direction of increasing σ values from 0 to 2 in increments of 0.25 units. The line for σ = 0 is the same for all second gases.
extremities of each log-normal distribution. Log-normal dis- in k, indicating that “shrinkage” need not be the only factor
tributions are not the only distributions seen clinically,16–20 involved in producing the second gas effect. We have shown
but we believe the choice is not particularly relevant to the that k is affected by the degree of V! / Q! mismatch and that
conclusions reached here, which depend primarily on the the effect of V! / Q! mismatch on k differs in the gas phase
application of the Fick principle. and in blood. With no V! / Q! mismatch, σ = 0 and k = 1
Equation 1 provides an explanation of how the second in both phases. This is the situation that corresponds to the
gas effect can continue, even after nitrous oxide uptake classical volume “shrinkage” explanation of the second gas
decreases to levels seen in the “maintenance phase.” The effect. As V! / Q! mismatch increases, σ increases. In blood,
( )
term k V!I − V!A / V!A is responsible for the second gas as σ increases, k also increases and is greater than 1. In the
gas phase, as σ increases, k decreases and is less than 1. The
( )
effect. Here the difference V!I − V!A is produced by nitrous term k may be thought of as a scaling factor, controlling the
oxide. A decrease in this difference as nitrous oxide uptake
decreases can be compensated for by a simultaneous increase
( )
degree to which the volume “shrinkage,” V!I − V!A / V!A , is
expressed.
Our findings are consistent with the experimental clini- exist. This is consistent with most previous clinical and
cal findings of Peyton et al.,10 that the magnitude of the theoretical studies, including those of Peyton et al., who
second gas effect on arterial sevoflurane partial pressures measured the second gas effect on arterial sevoflurane and
was significantly greater than that measured simultaneously oxygen partial pressures in anesthetized patients.5,10 The aug-
on end-tidal partial pressures in patients immediately after mentation to blood desflurane partial pressures at 20 to 30
induction of anesthesia and for 30 min afterward. However, min predicted by the model is in the order of 10 to 15%
these authors did not attribute the finding to a concomitant with moderate degrees of V! / Q! scatter. This compares to a
diminution in second gas effect in the gas phase. Our model mean 12% increase in arterial sevoflurane partial pressures
provides a more precise explanation for their findings. While at this time point in ventilated patients measured by Peyton
effects in the gas phase are most easily detected using end- et al.10 The lesser augmentation in gas phase desflurane con-
tidal gas concentration monitoring by the infrared analyzers centrations predicted here is also broadly consistent with the
available in most operating theaters today, augmentation in findings of Taheri and Eger, and those of Hendrickx et al.
the blood phase is not measured in normal clinical practice for end-expired sevoflurane concentrations, in anesthetized
but is important because it most directly affects anesthetic patients.4,5
partial pressures in blood and therefore depth of anesthe- Modeling and predicting the effects on respiratory gases
sia. An obvious clinical implication of these findings is is more complex and was not part of the scope of this study.
that minimum alveolar concentration (MAC) calculations In their previous studies using multicompartment model-
based on end-tidal concentration measurements may well ing of lung gas exchange, Peyton et al. modeled oxygen as
underestimate the depth of anesthesia when nitrous oxide a second gas,6–8 and demonstrated significant elevation of
is supplemented with a volatile agent. This represents a new predicted partial pressures of oxygen in arterial blood with
factor that should be taken into consideration in addition moderate degrees of V! / Q! scatter, at 45 min or more into
to other relevant factors22–24 when interpreting an end-tidal inhalational anesthesia with 70% nitrous oxide as the first
MAC reading during anesthesia. gas. This prediction was confirmed in a subsequent clinical
Figure 2 illustrates the predicted effects of typical study.9 In attempting to mimic oxygen uptake as closely as
increases in the degree of V! / Q! scatter on the augmenta- possible, these authors introduced the nonlinear binding of
tion ratio. We can treat the line for the homogenous lung, oxygen to hemoglobin, shunt formation, hypoxic vasocon-
i.e., the line for which σ = 0, as the first line in the series. This striction, absorption atelectasis, carbon dioxide elimination
line is identical in the upper (gas phase) and lower (blood and carriage, and acid-base status. Each of these factors is a
phase) diagrams in figure 2, as expected for a homogeneous potential confounder in interpretation of their findings of
“single compartment” lung. As steady state conditions are the relationship of V! / Q! scatter to the second gas effect. In
approached for the first gas, augmentation in a homogenous contrast, we have chosen to eliminate all but the essential
lung is independent of the second gas solubility in blood. components in this study of the second gas effect, permitting
It depends only on the ratio of the uptake of nitrous oxide explicit mathematical expression of the underlying features,
( )
to the alveolar ventilation, which is equal to V!I − V!A / V!A . once a particular mathematical distribution of V! / Q! ratios
Once ventilation and perfusion are no longer perfectly has been assumed. It remains to be seen to what degree these
matched, solubility becomes the important factor. other factors affect the predictions made here and should be
It is evident from figure 2 that the second gas effect in the subject of further work in the field.
blood and its augmentation by increasing V! / Q! scatter is In summary, modeling of ventilation-perfusion inhomoge-
greater for the least soluble second gas, desflurane. This neity confirms that the second gas effect is greater in blood
finding is remarkable, as lung modeling in the absence of than in expired gas. As the degree of mismatch increases, the
nitrous oxide predicts that increasing V! / Q! scatter, with magnitude of the second gas effect increases in blood but
a reduction in effective alveolar ventilation, impairs alve- decreases in expired gas. The change is most pronounced in
olar-capillary gas exchange more severely for less soluble blood for the less soluble volatile agents in current use. Clini-
gases,15 a prediction recently confirmed in an animal study cally, this means that MAC readings may well underestimate
of desflurane or isoflurane anesthesia using methacholine the depth of anesthesia when nitrous oxide is supplemented
inhalation to induce increased V! / Q! inhomogeneity.25 with a volatile agent.
This would suggest that inclusion of nitrous oxide might
maintain the efficiency of the lung in exchanging less sol- Research Support
uble modern volatile agents in the face of V! / Q! scatter Supported by an Australian Government Research Training
induced by anesthesia itself. Program Fees Offset (to Dr. Korman) and by grant Nos. U01-
Our mathematical modeling suggests, moreover, that the HL122199 and P01-GM066730 from the National Institutes
second gas effect in blood for less soluble modern agents is of Health (Bethesda, Maryland; to Dr. Dash).
expected to remain significant, although progressively reduc-
ing, into the “maintenance phase” of inhalational anesthesia Competing Interests
with nitrous oxide, where typical degrees of V! / Q! scatter The authors declare no competing interests.
Correspondence ⎛ σ⎞
2
1 − ⎝⎜X + 2 ⎠⎟ 2
Address correspondence to Dr. Korman: Department of An- q! = e (1.9)
aesthesia and Pain Medicine, Royal Perth Hospital, Perth, 2π
Western Australia 6000, Australia. ben@korman.com.au.
This article may be accessed for personal use at no charge Equations 1.8 and 1.9 are of the form of a standard normal
through the Journal Web site, www.anesthesiology.org. distribution and have been used to plot figure 1.
These forms of the density function for the log-normal
Appendix 1: The Log-normal Distribution distributions are very useful in equilibrating gas and blood
Colburn et al.15 derive the following probability density compartment by compartment, in order to obtain flow-
functions for ventilation and blood flow from first weighted outputs in gas and blood. The x-axis is divided
principles: into N compartments of equal width. This involves N+1
points. At least 99.7% of cases26 will always be covered if
V! − x 2 the subdivision of the x-axis begins at –σ/2 – 3 and ends at
V! = A e
2
(1.1) σ/2 + 3. The width of each interval is then (σ + 6)/N. When
2π
calculating the contribution of each point to the output, it
is necessary to apply the trapezoidal rule so that the first and
Q!
Q! = T e ( )
2
− x −σ / 2
(1.2) last points carry half the weight of the other points.27 If this
2π is not done, an error will be introduced that becomes more
significant as the number of compartments is reduced. In
Here V!A is the alveolar ventilation and Q!T the total pul- determining the weight attached to the expired ventilation
monary blood flow. Letting v! = V! V!A and q! = Q! Q!T , the from each compartment, it is necessary to multiply v! by
fractional or normalized ventilation and blood flow respec- the expired alveolar ventilation, V!A . Similarly, q! must be
tively, we have multiplied by Q!T .
1 − x2 2
Appendix 2: Derivation of Equation 1
v! = e (1.3)
2π
Washin Ratio
1 − ( x −σ )2 / 2 The proportion of nitrous oxide washin completed is expressed
q! =
2π
e (1.4)
(
as a washin ratio (WR), where WR = 1 − V!N O V!IFIN O .
2 2 )
This ratio is used as an independent variable on the x-axis.
Dividing equation 1.3 by equation 1.4: Note that when V!N2 O = 0, WR = 1 and nitrous oxide uptake is
complete; when V!N2 O = V!IFIN 2O , its maximum possible value,
v!/q! = e ⎣
{ }{ }
⎡ ( x −σ )2 / 2 − x 2 2 ⎤
⎦ (1.5) WR = 0, a situation that can only apply at the commence-
ment of anesthesia (see table A2.1). Since, for simplicity, the
i.e., equations we use are only applicable in the steady state, we
restrict our attention to values of V!N2 O for 0.9 ≤ WR ≤ 1.0,
i.e., when washin of nitrous oxide is 90 to 100% complete
v!/q! = e
(σ 2
)
− 2σ x 2 (1.6)
and steady state equations represent a satisfactory approxima-
tion. Applying Severinghaus’s equation for the rate of uptake
of nitrous oxide,13 this should occur within 10 to 15 min
Taking natural logarithms of both sides and solving for x,
from the start of induction, at which time the rate of change
we obtain
in the uptake of nitrous oxide has declined to less than 5% of
its current rate of uptake.
σ 1
x= − log ( v!/q! ) (1.7)
2 σ
Basic Equations
1 As the steady state equations governing the exchange of the
Letting X =
σ
( )
log v! q! and substituting in equations 1.3 first gas in a single lung unit,* we take
and 1.4, we obtain V!IFI − V!AF = Q! λ( F − Fv ) (2.1)
2
⎛ σ⎞
1 − ⎝⎜X − 2 ⎠⎟ 2 and
v! = e (1.8)
2π (1 − F )V!A = (1 − FI )V!I (2.2)
Table A2.1. Nitrous Oxide Washin Ratio ⎛1 − F ⎞ !

V!I = ⎜ VA
⎝1 − FI ⎟
⎠ (2.5)
Stage of Anesthesia Washin Ratio N2O Uptake FA/FI
Start of washin 0 V̇
IFIN2O 0 Substituting for V!I in equation 2.1, we may solve this equa-
Washin complete 1 0 1 tion for F.
FA/FI = alveolar/inspired concentration ratio.
where V!I is incoming alveolar ventilation per unit time (the V!AFI + λQ! (1 − FI )Fv
F= (2.6)
“inflow”), V!A is outgoing alveolar ventilation per unit time V!A + λQ! (1 − FI )
(the “outflow”), Q! is blood flow per unit time, FI is concen- Second Gas Effect in Blood Phase. We denote the uptake of
tration of the first gas in the dry fraction of the inflow, Fv is the first gas, nitrous oxide, by the symbol VN! 2O , the uptake
concentration of the first gas in the dry fraction of a sample of the second gas in the absence of the first gas, by V!SG ,
of gas equilibrated with the incoming mixed venous blood, and the uptake of the second gas in the presence of first gas,
F is concentration of the first gas in the dry fraction of the by V!SG / FG . VN! 2O is given by the left-hand side of equation
outflow, which is also in equilibrium with the outgoing arte-
2.1; V!SG is given by the left-hand side of equation 2.3 when
rial blood, and λ is the Ostwald solubility of the first gas in
inspired alveolar ventilation equals V!A ; and V!SG / FG is given
blood at 37°C.
by the left-hand side of equation 2.3 with inspired alveolar
To this system we now add a second soluble gas, denoted
ventilation determined from equation 2.2. Using the super-
by the subscript SG to allow it to be distinguished from the
script hl for a homogenous lung,
first gas. The second gas also obeys equation 2.1:
V!IFISG − V!AFSG = Q! λSG ( FSG − FvSG ) λV!A( FI − Fv )

(2.3) V!N hlO = (2.7)
2
λ(1 − FI ) + (V!A Q! )
Based on the conservation of mass, equations 2.1 and 2.3 state
that the amounts of the first and second gases leaving the gas
phase must equal the amounts of these gases gained by the ⎛ V!A λSG FI SG ⎞
V!SGhl = ⎜ ⎟ (2.8)
blood phase. The remaining gas is treated as if it were completely ⎝ λSG + (V!A Q! )⎠
insoluble, and merely serves as a “vehicle” for delivering the first
and second gases. Equation 2.2 expresses the conservation of
the remaining gas. The second gas is assumed to be present in ⎛ V!A λSG FI SG ⎞ ⎛ 1 − F ⎞
V!SGhl/FG = ⎜ ⎟⎜ ⎟ (2.9)
such a low concentration that its contribution to the net volume ⎝ λSG + (V!A Q! )⎠ ⎝1 − FI ⎠
change may be ignored. Thus, it does not appear in equation
2.2. For convenience, we set FvSG to zero, effectively excluding Note that equation 2.7 was derived by Colburn et al.15
recirculation. This maximizes the uptake of the second gas, mak- Except for the appearance of the factor (1 – F)/(1 – F I)
ing the detection of the second gas effect easier. Clinically, this is and the omission of the Fv term, equation 2.9 is similar
equivalent to the administration of minute traces of second gas to equation 2.7. Because F ISG is much smaller than 1, it
at intermittent intervals. From equation 2.3 we then have has been omitted from the denominator of Eqs. 2.8 and
2.9. From Eqs. 2.8 and 2.9, it is trivial to find that the
augmentation ratio in blood, AR(blood) = V!SGhl/FG / VSG! hl =
V!IFISG
FSG = (2.4) (1 – F)/(1 – F I).
V!A + Q! λSG
Second Gas Effect in Gas Phase. Substituting for inspired
alveolar ventilation from equation 2.2 in equation 2.3, we
Homogenous Lung obtain an expression for FSG / FG :
This case is obtained by solving equation 2.2 for V!I and
eliminating this variable from all equations. ⎛ (V!A / Q! )FISG ⎞ ⎛ 1 − F ⎞
FSGhl / FG = ⎜ ⎜ ⎟ (2.10)
⎝(V!A / Q! ) + λSG ⎟
⎠ ⎝1 − FI ⎠
*Colburn et al.15 present their equations 1 and 2 in exactly
the same form as our equations 2.1 and 2.2 except that frac- When no first gas is present, inspired alveolar ventilation
tional concentrations in dry gas are replaced by partial pres-
sures. Their equation 2 is then clearly incorrect and should read
equals expired alveolar ventilation and (1 – F)/(1 – FI) = 1,
( PB - PH 2 O - P )V!A = ( PB - PH 2 O - PI )V!I If we then divide this equa- and hence the expression for FSGhl can be easily derived from
( )
tion and their equation 1 by PB - PH 2 O , we obtain our equations equation 2.10. Thus, the augmentation ratio in the gas
2.1 and 2.2, which have exactly the same form as their original phase, AR(gas) = FSGhl/FG/ FSGhl = (1 – F)/(1 – FI).
equations except that fractional concentrations replace the partial We can conclude that in a homogenous lung with con-
pressures. This means that all of their subsequent results are valid.
The correct form of equation 2.2 was used in a subsequent publica- stant outflow, the augmentation ratios in the gas phase and
tion by two of the three original authors.28 in the blood phase are both equal to (1 – F)/(1 – FI) so that
the magnitude of the second gas effect is identical in both ⎡ λ( FI − Fv ) ⎤⎡ I ⎤

phases and, after substituting the solution for F (obtained AR(blood) = 1 + ⎢ ⎥⎢ − 1⎥ (2.16)
from equation 2.1), is given by ⎣λSG − λ(1 − FI ) ⎥
⎢ ⎦⎣I SG ⎦
(V!A / Q! ) + λ(1 − Fv ) where

AR(gas) = AR(blood) = (2.11)
(V!A / Q! ) + λ(1 − FI ) I =
1 ∞ dx
∫
2π − ∞ 2
( ) 2
λ (1 − FI )exp{ x / 2 } + V!0 Q!0 exp{ (x − σ ) / 2 }
Nonhomogenous Lung n ⎛ (V! j V!0 ) ⎞
Consider now a nonhomogenous lung model, denoted by ≡ ∑ ⎜
j =1 ⎝ λ (1− FI )+ (V! j Q! j )⎟
⎠
the superscript nhl, composed of n functional units. Let Vj!
∞
and Q! j be the alveolar ventilation and blood flow, respec- I SG =
1
∫
dx
tively, to the jth compartment for 1 ≤ j ≤ n. Set V!0 = V!1 + ( )
2π − ∞ λ exp{ x / 2 } + V! Q! exp{ (x − σ )2/ 2 }
2
0 0
SG
…. + V!n and Q!0 = Q!1 + ….. + Q!n .
n ⎛ (V! j V!0 ) ⎞
Second Gas Effect in Blood Phase. By analogy with equa- ≡ ∑ ⎜
j =1 ⎝ λSG + (V! j Q! j )⎟
⎠
tion 2.9, the net uptake of the second gas in the presence of
the first gas in the nonhomogenous lung, V!SGnhl/ FG , is given by (2.17)
Using the substitution Fv = FI*(Fv/FI), we can now deter-
n
⎛ Vj! λ FI ⎞ ⎛1 − F j ⎞
V!SGnhl/ FG = ∑ ⎜ λ +SG(V! SG (2.12) mine the augmentation ratio in blood, AR(blood), for a
⎝ SG j Q!j ) ⎟
⎠⎜⎝1 − FI ⎟
⎠
j =1 range of values of λSG, σ, and Fv/FI (representing different
stages of nitrous oxide equilibration). However, Peyton et al.
By analogy with equation 2.8, the net uptake of the second
actually use the nitrous oxide uptake, V!N O , as an input vari-
gas in the absence of the first gas in the nonhomogeneous 2
able.6–8 This enables us to eliminate the term (FI – Fv), as
lung, V!SGnhl , is given by
described in equation 2.19, below. Using equation 2.7, the
n
⎛ Vj! λ FI ⎞ result derived by Colburn et al.15 for the uptake of nitrous
V!SGnhl = ∑ ⎜ λ +SG(V! SG
j =1
⎝ SG j Q!j ) ⎟
⎠ (2.13) oxide in a homogenous lung and summing the uptake in
each compartment to obtain the total uptake, we obtain the
In the case of a nonhomogenous lung, the augmentation following equation:
ratio for blood is therefore given by
n λV!j ( FI − Fv )
n
⎛ V! ⎞ ⎛1 − Fj ⎞ V!N 2nhl =∑ (2.18)
∑ ⎜⎝ λSG +j( V!j Q!j ) ⎟
⎠⎜⎝1 − FI ⎟
⎠
O ! !
j =1 λ (1 − FI ) + (Vj Q j )
j =1
AR(blood) = (2.14) which, by analogy, produces the following expression for
n
⎛ V! ⎞
∑ ⎜⎝ λSG +j( V!j Q!j ) ⎟
⎠
(FI – Fv) if we assume a log-normal distribution of ventila-
j =1
tion-perfusion ratios:
Applying equation 2.6 to the jth compartment and simplify- V!N nhl
ing, we obtain the following expression for the augmenta- ( FI − Fv )= 2O
(2.19)
λ V!0 I
tion ratio in blood:
The final expression relating the second gas effect to the
⎡ λ( FI − Fv ) ⎤ uptake of nitrous oxide and the degree of ventilation-perfu-
AR(blood) = 1 + ⎢ ⎥ sion mismatch in a nonhomogenous lung is therefore
⎣λSG − λ(1 − FI ) ⎥
⎢ ⎦
⎡n ⎛ (V!j V!0 ) ⎞ ⎤ ⎛ ⎞ ⎛V!Nnhl ⎞
⎢∑ ⎜
(2.15) 1
λ (1− FI )+ (V!j Q!j ) ⎟⎠ ⎥
2O
⎝ AR(blood) = 1 + ⎜ ⎟ ⎜ ! ⎟
⎢j =1 − 1⎥ ⎝ λ SG − λ (1 − FI ) ⎠ ⎝ V0 ⎠ (2.20)
⎢ n ⎛ (V!j V!0 ) ⎞ ⎥
⎢ ∑⎜ λ + !j Q!j ) ⎟ ⎥ ⎛ 1 1⎞
⎣ j =1 ⎝ SG ⎠
(V −
⎦ ⎜
⎝ I SG I ⎟
⎠
Note here that with 70% nitrous oxide (i.e., FI = 0.7) for We use WR, the previously defined washin ratio, as a variable
which λ = 0.47, the expression [ λSG − λ(1 − FI )] is positive on the x-axis. Note that given the assumption that only the first
for all λSG > 0.141, which includes all the volatile anesthetic
gas, nitrous oxide, contributes to changes in gas volume, V!I =
agents in current use. Assuming now that these series are
both applied using a log-normal distribution of V! / Q! with V!A + V!N2 O .
log SD of ventilation-perfusion mismatch of σV! − σ Q! , we Second Gas Effect in Gas Phase. Whereas the second gas
replace each series with its equivalent Lebesgue-Stieltjes inte- is absorbed by the same volume of blood whether or not
gral as described by Colburn et al.15 Thus, we have nitrous oxide is present, this is not the case in the gas phase.
Here the second gas is contained in a smaller volume whose Physiological modeling of gas exchange. J Appl Physiol
(1985) 2001; 91:17–25
size depends on the rate of nitrous oxide uptake. The equa-
9. Peyton PJ, Horriat M, Robinson GJ, Pierce R, Thompson BR:
tions describing this phase are therefore more complicated. Magnitude of the second gas effect on arterial sevoflurane
Only the final result is given here. partial pressure. ANESTHESIOLOGY 2008; 108:381–7
10. Peyton PJ, Stuart-Andrews C, Deo K, Strahan F, Robinson
⎛ 1 ⎞ ⎛V!NnhlO ⎞ GJ, Thompson BR, Pierce R: Persisting concentrating and
AR(gas) = 1 + ⎜ 2 second gas effects on oxygenation during N2O anaesthesia.
⎟⎜ ! ⎟
λ(
⎝ SG − λ (1 − FI ) I ⎠ ⎝) V 0 ⎠
(2.21) 11.
Anaesthesia 2006; 61:322–9
Severinghaus JW, Hendrickx JF, Carette R, Lemmens HJ, De
⎛ λ SG I SG − λ (1 − FI ) I ⎞ Wolf AM: Can large volume N2O uptake alone explain the
⎜ 1 − λ SG I SG
⎟ second gas effect? Br J Anaesth 2006; 97:262; author reply
⎝ ⎠ 262–3
12. Magnusson L, Spahn DR: New concepts of atelectasis during
AR(gas) is the augmentation ratio in the gas phase. Note that general anaesthesia. Br J Anaesth 2003; 91:61–72
equations 2.20 and 2.21 may both be simplified to the follow- 13. Severinghaus JW: The rate of uptake of nitrous oxide in man.
J Clin Invest 1954; 33:1183–9
ing approximate form:
14. West JB, Wagner P: Pulmonary gas exchange, Bioengineering
AR = 1 + k(1 − WR ) (2.22) aspects of the lung. Edited by West JB. New York, Marcel
Dekker Inc., Publishers, 1977, pp 378–82
where k is a constant that differs for gas and blood and
15. Colburn WE, Evans JW, West JB: Analysis of effect of the
incorporates the effect of σ. When the augmentation ratio solubility on gas exchange in nonhomogenous lungs. J Appl
equals 1, there is no second gas effect, so when the second Physiol 1974; 37:547–51
gas effect occurs, we can write second gas effect, SGE = 16. Rehder K, Knopp TJ, Sessler AD, Didier EP: Ventilation-
perfusion relationship in young healthy awake and anesthe-
k(1 – WR). This indicates how the second gas effect can tized-paralyzed man. J Appl Physiol Respir Environ Exerc
be both a function of the degree of nitrous oxide uptake Physiol 1979; 47:745–53
(1 – WR) and the degree of ventilation-perfusion mismatch 17. Dueck R, Young I, Clausen J, Wagner PD: Altered distribution
(which appears in the term k along with the solubility of of pulmonary ventilation and blood flow following induction
of inhalation anesthesia. ANESTHESIOLOGY 1980; 52:113–25
the first and second gases and the inspired concentration
18. Lundh R, Hedenstierna G: Ventilation-perfusion relationships
of the first gas). during anaesthesia and abdominal surgery. Acta Anaesthesiol
Perhaps the most useful form of equation 2.21 involves Scand 1983; 27:167–73
the substitution of V!I − V!A for V!Nnhl
2O
. This gives rise to the 19. Lundh R, Johansson H: Alveolar stability during anaes-
thesia for reconstructive vascular surgery in the leg. Acta
following equation: Anaesthesiol Scand 1983; 27:26–34
AR = 1 + k
(V!I − V!A ) (2.23)
20. Lundh R, Hedenstierna G: Ventilation-perfusion relation-
ships during halothane anaesthesia and mechanical ventila-
V!A tion. Effects of varying inspired oxygen concentration. Acta
Anaesthesiol Scand 1984; 28:191–8
References 21. Dantzker D, Wagner P, West JB: Instability of lung units with
low VA/Q ratios during O2 breathing. J Appl Physiol 1975;
1. Epstein RM, Rackow H, Salanitre E, Wolf GL: Influence of the 38:886–95
concentration effect on the uptake of anesthetic mixtures: 22. Aranake A, Mashour GA, Avidan MS: Minimum alveolar con-
The second gas effect. ANESTHESIOLOGY 1964; 25:364–71 centration: Ongoing relevance and clinical utility. Anaesthesia
2. Stoelting RK, Eger EI 2nd: An additional explanation for the 2013; 68:512–22
second gas effect: A concentrating effect. ANESTHESIOLOGY 23. Eger EI 2nd, Tang M, Liao M, Laster MJ, Solt K, Flood P,
1969; 30:273–7 Jenkins A, Raines D, Hendrickx JF, Shafer SL, Yasumasa T,
3. Korman B, Mapleson WW: Concentration and second gas Sonner JM: Inhaled anesthetics do not combine to produce
effects: Can the accepted explanation be improved? Br J synergistic effects regarding minimum alveolar anesthetic
Anaesth 1997; 78:618–25 concentration in rats. Anesth Analg 2008; 107:479–85
4. Taheri S, Eger EI 2nd: A demonstration of the concentration 24. Katoh T, Ikeda K, Bito H: Does nitrous oxide antago-
and second gas effects in humans anesthetized with nitrous nize sevoflurane-induced hypnosis? Br J Anaesth 1997;
oxide and desflurane. Anesth Analg 1999; 89:774–80 79:465–8
5. Hendrickx JF, Carette R, Lemmens HJ, De Wolf AM: Large 25. Kretzschmar M, Kozian A, Baumgardner JE, Borges JB,
volume N2O uptake alone does not explain the second gas Hedenstierna G, Larsson A, Hachenberg T, Schilling T: Effect
effect of N2O on sevoflurane during constant inspired venti- of bronchoconstriction-induced ventilation-perfusion mis-
lation. Br J Anaesth 2006; 96:391–5 match on uptake and elimination of isoflurane and desflu-
6. Peyton PJ, Robinson GJ, Thompson B: Ventilation-perfusion rane. ANESTHESIOLOGY 2017; 127:800–12
inhomogeneity increases gas uptake: Theoretical modeling 26. Spiegel MR: Schaum's Outline of Theory and Problems of
of gas exchange. J Appl Physiol (1985) 2001; 91:3–9 Statistics in SI units, 1st edition. New York, McGraw Hill,
7. Peyton PJ, Robinson GJ, Thompson B: Ventilation-perfusion 1972, p 123
inhomogeneity increases gas uptake in anesthesia: 27. Mathews JH, Fink KD: Numerical Methods using Matlab, 3rd
Computer modeling of gas exchange. J Appl Physiol (1985) edition. New Jersey, Prentice-Hall, 1999, p 354
2001; 91:10–6 28. Evans J, Wagner P, West J: Conditions for reduction of pulmo-
8. Peyton PJ, Robinson GJ, Thompson B: Effect of ventila- nary gas transfer by ventilation-perfusion inequality. J Appl
tion-perfusion inhomogeneity and N(2)O on oxygenation: Physiol 1974; 36:533–7
J Appl Physiol 126: 558–568, 2019.
First published December 6, 2018; doi:10.1152/japplphysiol.00689.2018.
RESEARCH ARTICLE
Effect of net gas volume changes on alveolar and arterial gas partial pressures
in the presence of ventilation-perfusion mismatch
Ben Korman,1 Ranjan K. Dash,2 and Philip J. Peyton3
1
Department of Anaesthesia and Pain Medicine, Royal Perth Hospital, Perth, Western Australia, Australia; 2Departments of
Biomedical Engineering and Physiology, Medical College of Wisconsin, Milwaukee, Wisconsin; and 3Anaesthesia,
Perioperative, and Pain Medicine Unit, Melbourne Medical School, University of Melbourne, Department of Anaesthesia,
Austin Health, Heidelberg, Victoria, Australia
Submitted 9 August 2018; accepted in final form 3 December 2018
Korman B, Dash RK, Peyton PJ. Effect of net gas volume pulmonary uptake of any gas with a linear dissociation curve
changes on alveolar and arterial gas partial pressures in the presence of (8). However, clinical investigations of the second gas effect
ventilation-perfusion mismatch. J Appl Physiol 126: 558 –568, 2019. (SGE) performed over the last decade have brought this prop-
First published December 6, 2018; doi:10.1152/japplphysiol.00689.
osition into question (9, 17). The SGE occurs when a soluble
2018.—The second gas effect (SGE) occurs when nitrous oxide enhances
the uptake of volatile anesthetics administered simultaneously. Recent first gas such as nitrous oxide (N2O) is delivered in high
work shows that the SGE is greater in blood than in the gas phase, that inspired concentrations during anesthesia (7, 25). The high
this is due to ventilation-perfusion mismatch, that as mismatch increases, inspired concentrations are associated with the transfer of large
the SGE increases in blood but is diminished in the gas phase, and that volumes of gas, of the order of 0.6 to 1 l/min from lungs to
these effects persist well into the period of nitrous oxide maintenance blood during the early stages of anesthesia (12, 22). Thereafter,
anesthesia. These modifications of the SGE are most pronounced with the the rate of N2O uptake falls progressively toward zero. Alve-
low soluble agents in current use. We investigate further the effect of net olar-capillary uptake of the first gas increases the alveolar
gas volume loss during nitrous oxide uptake on low concentrations of
other gases present using partial pressure-solubility diagrams. The steady-
concentrations of other gases present, accelerating their uptake.
state equations of gas exchange were solved assuming a log-normal These other gases include the commonly used volatile anes-
distribution of ventilation-perfusion ratios using Lebesgue-Stieltjes inte- thetic agents and are referred to individually and collectively as
gration. It was shown that under these conditions the classical partial second gases.
pressure-solubility diagram must be modified, that for currently used Until 2000, the most widely accepted explanation of the
volatile anesthetic agents the alveolar-arterial partial pressure difference SGE appealed to the contraction in volume associated with
is less than that predicted in the past, and that the alveolar-arterial partial N2O uptake as the factor responsible (12). The SGE was
pressure difference may even be reversed during uptake in the case of measured in expired gas and generally assumed to be of similar
highly insoluble gases such as sulfur hexafluoride. Comparing this with
the situation described previously for nitrogen in steady-state air breath-
magnitude in blood. More recently, it has been shown that the
ing, we show that for nitrogen, the direction of the alveolar-arterial effect is not only greater in blood but also continues to be
gradient is opposite to the direction of net gas volume movement. significant when N2O uptake falls to relatively low steady-state
Although gas uptake with ventilation-perfusion inequality exceeding that levels (9, 21). Based on computer simulations assuming a
when matching is optimal is shown to be possible, it is less likely than log-normal distribution of ventilation and blood flow, it was
alveolar-arterial partial pressure reversal. suggested that this newly recognized phenomenon is due to the
NEW & NOTEWORTHY Net uptake of gases administered with mismatch between ventilation and blood flow (V̇/Q̇ mismatch)
nitrous oxide may proceed against an alveolar-arterial partial pressure that occurs routinely following the induction of anesthesia
gradient. The alveolar-arterial gradient for nitrogen in the steady-state (18 –20).
breathing air depends not only on the existence of a distribution of In an earlier study, West et al. (26) obtained partial pressure-
ventilation-perfusion ratios in the lung but also on the presence of a solubility diagrams in the case of a single soluble inert gas
net change in gas volume and is opposite in direction to the direction present in the inspired gas mixture together with an insoluble
of net gas volume uptake.
gas that acts solely as a vehicle for delivering the soluble inert
alveolar-arterial partial pressure gradient; anesthetic uptake; mathe- gas. Their study focused on the elimination from the lung of
matical modeling; second gas effect; ventilation-perfusion mismatch a gas supplied at constant partial pressure in mixed venous
blood. Colburn et al. (3) explored further the exchange of inert
gases in such a system. They concluded that mismatching of
INTRODUCTION ventilation and blood flow must reduce the pulmonary uptake
of any gas with a linear dissociation curve, a conclusion stated
A long-held tenet of respiratory physiology is the belief that explicitly by Evans et al. (8), in a separate publication. How-
mismatching of ventilation and blood flow must reduce the ever, these authors did not consider possible effects associated
with net gas volume changes during gas exchange. In this
Address for reprint requests and other correspondence: B. Korman, Dept. of
paper, we use N2O uptake as a tool to investigate the effect of
Anaesthesia and Pain Medicine, Royal Perth Hospital, Perth, WA 6000, net gas volume changes on the partial pressure-solubility dia-
Australia (e-mail: ben@korman.com.au). gram and its effect on gas uptake in nonhomogenous lungs.
558 8750-7587/19 Copyright © 2019 the American Physiological Society http://www.jappl.org
Downloaded from www.physiology.org/journal/jappl at Medical Col of Wisconsin (141.106.122.102) on March 8, 2019.
VENTILATION-PERFUSION MISMATCH AND THE SECOND GAS EFFECT 559
METHODS ⌬V̇/V̇A. Note that (V̇I ⫺ V̇A)/V̇A is equal to (V̇I/V̇A) ⫺ 1 and is
therefore also the relative volume change associated with N2O uptake.
Consider gases with linear dissociation curves, commonly referred For convenience, we set Fv៮ in Eq. 6 equal to 0. Alveolar ventilation
to as inert gases, being exchanged under steady-state conditions. If V̇A is set at 4 l/min, and pulmonary blood flow Q̇ is set at 5 l/min.
such a gas is delivered in the inspired gas mixture at a rate of V̇IFI and Partial pressure-solubility relationship for soluble second gas.
removed at a rate of V̇AFA in a homogenous lung of given V̇A/Q̇, then Consider now a second soluble gas present in such low concentrations
the uptake of the gas is V̇IFI ⫺ V̇AFA. This must equal the uptake by that it has no impact on Eq. 2. This will also obey Eq. 1 in each gas
blood, which is given by ␭Q̇(Fa ⫺ Fv៮ ), where Fa and Fv៮ are equal to exchanging unit. Assuming a mixed venous partial pressure of zero
Pa ⁄ 共PB ⫺ PH2O兲 and Pv៮ / 共PB ⫺ PH2O兲 respectively, so that for this agent, expressions may be obtained for the mixed alveolar and
mixed pulmonary end-capillary partial pressures as follows:
V̇IPI ⫺ V̇APA ⫽ ␭Q̇(Pa ⫺ PV៮ ). (1)
PA៮ N 共V̇ ⁄ Q̇ j 兲
For the diluent gas, which is treated as if it were completely ⫽ 兺
Ij
共V̇ ⁄ V̇A 兲 (7)
共V̇ 兲
Aj
insoluble, we have PI j⫽1 ⁄ Q̇ j ⫹ ␭
Aj
V̇I共1 ⫺ FI兲 ⫽ V̇A共1 ⫺ FA兲 . (2) and

Assuming complete equilibration between gas and blood so that P៮a N 共V̇ ⁄ Q̇ j 兲
Pa ⫽ PA, these equations give rise to the following general equation ⫽ 兺
Ij
共Q̇ ⁄ Q̇ 兲 . (8)
共V̇ 兲
t
for the alveolar and arterial partial pressures of the inert gas: PI j⫽1 ⁄ Q̇ j ⫹ ␭
j
Aj
共V̇A ⁄ Q̇兲PI ⫹ ␭共1 ⫺ FI兲PV៮ Here, ␭ is the solubility of the second gas, the subscript j refers to
Pa ⫽ PA ⫽ . (3) the jth compartment, and the symbol Q̇t refers to the total pulmonary
共V̇A ⁄ Q̇兲 ⫹ ␭共1 ⫺ FI兲 blood flow. If a log-normal distribution of ventilation-perfusion ratios
Given that V̇I/Q̇ ⫽ (V̇I/V̇A) ⫻ (V̇A/Q̇), Eq. 2 provides the following is assumed, Eqs. 7 and 8 may also be converted to their equivalent
expression for V̇I/Q̇: Lebesgue-Stieltjes forms and then solved for PA៮ /PI and Pa៮ /PI respec-
tively. Graphs may now be drawn for each of these parameters as a
共V̇A ⁄ Q̇兲 ⫹ ␭共1 ⫺ FV៮ 兲 function of ␭ for different values of ␴ and ⌬V̇.
共V̇I ⁄ Q̇兲 ⫽ 共V̇A ⁄ Q̇兲 . (4) The augmentation-solubility relationship. A useful measure of the
共V̇A ⁄ Q̇兲 ⫹ ␭共1 ⫺ FI兲 magnitude of the SGE is the augmentation ratio (AR), which is
defined as the partial pressure of the second gas in the relevant phase
Inspired alveolar ventilation with a log-normal distribution. In a (gas or blood) in the presence of N2O divided by the partial pressure
lung model consisting of N parallel compartments, the inspired the second gas would have in that phase in the absence of N2O, with
ventilation is the sum of the individual values for each compartment: all other factors being held constant (7, 11). When AR ⫽ 1, there is no
N SGE, so when the SGE does occur, its magnitude is given by the
V̇I ⫽ 兺
j⫽1
V̇I .
j
(5) amount by which AR exceeds unity. We have previously shown that
when the distribution of ventilation-perfusion ratios in the lung is log
When dealing with a log-normal distribution of expired alveolar normal, the following relationship applies:
ventilation-perfusion ratios, Colburn et al. (3) have shown that if ␴V̇
is the log standard deviation of ventilation and ␴Q̇ is the log standard ⌬V̇
AR ⫽ 1 ⫹ k . (9)
deviation of blood flow, then ␴ ⫽ |␴V̇ ⫺ ␴Q̇| specifies the degree of
V̇A
mismatch between ventilation and blood flow. In awake healthy
adults, ␴ varies from 0.25 to 1.0; worsening of V̇/Q̇ matching is seen Here, k is a variable that depends primarily on the solubility of the
in disease states and is also known to occur soon after induction of second gas in blood (␭) and the degree of mismatch between venti-
anesthesia, with typical values for ␴ of 0.75–1.5 in healthy adults (4, lation and blood flow (␴) (10, 11). Equation 9 explains how it is
13–15, 24). With a log-normal distribution of V̇/Q̇ ratios, Eq. 5 may possible for the degree of ventilation-perfusion mismatch to influence
then be expressed as the equivalent Lebesgue-Stieljes integral and the magnitude of the SGE provided that ⌬V̇/V̇A, the relative volume
becomes change associated with N2O uptake, remains greater than zero. Aug-
再
mentation ratios may be obtained for the gas phase and for blood by
V̇A
兰 dividing PA៮ /PI and Pa៮ /PI at any given value of ⌬V̇ with their equiv-
⬁
V̇I ⫽ exp(⫺x2 ⁄ 2)
⫺⬁ alent values when ⌬V̇ ⫽ 0. In each case, all other variables must be
兹 2␲
冎
held constant. AR(gas) and AR(blood) may then be graphed as
␭共FI ⫺ FV៮ 兲 functions of ␭.
⫹ dx. (6)
共V̇A ⁄ Q̇兲exp(x ⫺ ␴)2 ⁄ 2 ⫹ ␭共1 ⫺ FI兲exp(x2 ⁄ 2) Alveolar-arterial partial pressure differences. An estimate of the
A-a difference may be obtained by multiplying the difference
The left-hand term in the integrand represents the log-normal {PA៮ /PI ⫺ Pa៮ /PI} by the partial pressure of the second gas in the dry
distribution of expired ventilation-perfusion ratios. When the first gas inspired gas mixture. For example, when 1% sevoflurane is present in
is nitrous oxide, ␭ ⫽ 0.47 and the right-hand term gives the uptake of the dry inspired mixture, it is necessary to multiply the difference by
nitrous oxide. Once V̇A, Q̇, ␴, and Fv៮ are specified, Eq. 6 provides a 7.13 to obtain an estimate of the A-a difference in mmHg.
one-to-one relationship between V̇I and FI and hence, allows the The uptake-solubility relationship. As an index of gas exchange,
calculation of the inspired concentration of nitrous oxide associated Colburn et al. (3) studied R(␭), the ratio of the gas exchange rate of
with the desired total inspired alveolar ventilation. Letting ⌬V̇ ⫽ a compartmental model over the exchange rate of the homogenous
(V̇I ⫺ V̇A), we may now determine the inspired concentration of N2O model with the same total expired alveolar ventilation and blood flow.
necessary to produce the required values of ⌬V̇ or ⌬V̇/V̇A, either of In their model, R(␭) is always ⬍1, indicating that gas exchange in a
which may be chosen as an independent variable for investigation, nonhomogenous lung is always less than in the equivalent homoge-
along with ␴, the degree of mismatch between ventilation and blood nous lung. We use their expression for R(␭) and apply it to an inert gas
flow. We use the terms “volume shrinkage,” “volume contraction,” present in the inspired gas mixture in low concentrations in the
and “net gas volume uptake” interchangeably when referring to presence of net volume changes generated by N2O uptake.
J Appl Physiol • doi:10.1152/japplphysiol.00689.2018 • www.jappl.org

560 VENTILATION-PERFUSION MISMATCH AND THE SECOND GAS EFFECT
All programs were written in MATLAB and solved using the fsolve noted by Stoelting and Eger (25). Until the work of Peyton et
nonlinear algebraic equations solver. The integral function was used al. (18 –20), no particular effect was ascribed to the addition of
to integrate improper integrals such as Eq. 6. The equations used to V̇/Q̇ mismatch. If the effects were merely additive, they would
generate the figures in the RESULTS and DISCUSSION sections are dis-
cussed further in APPENDIX 1 and APPENDIX 2. For individual figures,
produce the result shown in Fig. 1C.
values of ⌬V̇ were chosen to demonstrate differences when these Partial pressure-solubility diagrams in the presence of net
occur, while also lying within the range believed to occur clinically gas-volume uptake. Figure 2 shows what actually happens in
during N2O anesthesia. our model when a net gas volume reduction is combined with
V̇/Q̇ mismatch. It shows the effect of increasing ⌬V̇ from 0 to
RESULTS
600 ml/min on the alveolar and arterial partial pressures in a
We begin by drawing a partial pressure-solubility diagram of homogenous lung (i.e., ␴ ⫽ 0) and a lung with significant V̇/Q̇
the form described by West et al. (26). This is achieved in our mismatch (␴ ⫽ 1.5).
model by setting the inspired nitrous oxide concentration to With no volume change (⌬V̇ ⫽ 0), the graph of PA៮ /PI is
zero, so that with the very low inspired concentrations of the displaced to the right by V̇/Q̇ scatter, as in Fig. 1A. A similar
other soluble gas present, there is effectively no change in net displacement occurs when ⌬V̇ ⫽ 600. Now it is the increase in
gas volume. The resulting curves are shown in Fig. 1A. PA៮ /PI of the continuous blue line relative to the continuous
When there is no V̇/Q̇ mismatch, i.e., ␴ ⫽ 0, alveolar and black line that reflects the impaired uptake in the presence of an
arterial partial pressures are equal. This is indicated by the increased degree of mismatch. For arterial blood, when ⌬V̇ ⫽
black dash-dotted line in Fig. 1A. Moving along this line 0, the line for Pa៮ /PI is displaced to the left by V̇/Q̇ scatter, as in
toward the left end of the x-axis, we are dealing with less Fig. 1A. With ⌬V̇ ⫽ 600, the graph of Pa៮ /PI with ␴ ⫽ 1.5
soluble gases; very little is transferred to blood so that the (continuous red line) greatly exceeds that for ␴ ⫽ 0 (continu-
alveolar (and hence arterial) partial pressure approaches the ous black line) for much of its course, especially at low
partial pressure in the inspired gas mixture, and both PA៮ /PI and solubilities (␭ ⬍ 0.05). This reflects the increased uptake that
Pa៮ /PI approach 1. As we move along the black dash-dotted line occurs with significant gas volume shrinkage. The effect of the
in Fig. 1A toward the right end of the x-axis, we are dealing contraction in net gas volume on PA៮ /PI and Pa៮ /PI at each value
with more and more soluble gases; more and more of the of ␴ is obtained by comparing the dash-dotted line with the
alveolar gas is transferred to blood so that the alveolar (and solid line for each color. The greater magnitude of the SGE in
hence, the arterial) partial pressure approaches zero, and so do blood than in the gas phase is indicated by the larger area
the respective ratios PA៮ /PI and Pa៮ /PI. between the two red lines in Fig. 2B compared with the area
The presence of V̇/Q̇ mismatch acts as an impediment to gas between the two blue lines in Fig. 2A.
exchange; less gas is removed from the alveolar gas mixture,
Augmentation-solubility diagrams in the presence of net gas
producing a higher partial pressure in the gas phase and a lower
volume uptake. The augmentation or fractional increase in
partial pressure in blood. Thus, with ␴ ⫽ 1.5, the lines in Fig.
partial pressure of a gas due to gas volume shrinkage may be
1A for alveolar gas (blue dash-dotted line) and arterial blood
(red dash-dotted line) separate, producing an alveolar-arterial derived in Fig. 2 by dividing the value of PA៮ /PI or Pa៮ /PI when
partial pressure difference. At each end of the solubility scale, ⌬V̇ ⫽ 600 by its value when ⌬V̇ ⫽ 0 for the value of ␭
PA៮ /PI and Pa៮ /PI behave as before and approach 1 and 0, associated with that gas. In Fig. 3A, we have shown the
respectively. This is the classical description given by West et augmentation ratio with ␴ ⫽ 1.5 for gas (blue) and blood (red)
al. (26). when ⌬V̇ ⫽ 200 ml/min (dash-dotted lines) and 800 ml/min
Figure 1B depicts partial pressure-solubility diagrams in the (solid lines); augmentation, (and hence, the SGE) is greater at
absence of V̇/Q̇ mismatch, i.e., with ␴ ⫽ 0. The black dash- the higher value of ⌬V̇, an effect most noticeable as the
dotted line in Fig. 1B is the same as in Fig. 1A and represents solubility of the gas is decreased. In Fig. 3B, the augmentation
the situation where there is no net gas volume change (⌬V̇ ⫽ ratio is shown for gas and blood for a range of ␴ values with
0); the solid black line represents the situation with a net gas ⌬V̇ set at 100 ml/min. In the absence of V̇/Q̇ mismatch, the
volume loss of 600 ml/min (⌬V̇ ⫽ 600). The difference augmentation ratio is 1.025 (horizontal black line). Increasing
between the two lines reflects the magnitude of the second gas the degree of mismatch has opposite effects on the gas phase
effect. The effect is greater with gases of low solubility, a fact (blue lines) and blood (red lines). As ␴ increases, the SGE
Fig. 1. Alveolar and arterial partial pressures as a fraction of the inspired partial pressure vs. solubility expressed as log10(␭). A: PA៮ /PI (blue dash-dotted line)
and Pa៮ /PI (red dash-dotted line) with no net volume change i.e., ⌬V̇ ⫽ 0 and ␴ ⫽ 1.5 The black dash-dotted line represents the situation with ␴ ⫽ 0, in which
case PA៮ /PI and Pa៮ /PI are equal. B: no ventilation-perfusion mismatch, i.e., ␴ ⫽ 0; PA៮ /PI and Pa៮ /PI are equal, lower curve ⌬V̇ ⫽ 0 (black dash-dotted line) and
upper curve ⌬V̇ ⫽ 600 ml/min (black solid line). C: curves in A and B are summed for gas (blue) and blood (red).

Fig. 2. Alveolar and arterial partial pressures as a fraction of the inspired partial pressure vs. solubility expressed as log10(␭). A: PA៮ /PI with ␴ ⫽ 0, ⌬V̇ ⫽ 0 (black
dash-dotted line); ␴ ⫽ 0, ⌬V̇ ⫽ 600 ml/min (solid black line); ␴ ⫽ 1.5, ⌬V̇ ⫽ 0 (blue dash-dotted line); ␴ ⫽ 1.5, ⌬V̇ ⫽ 600 ml/min (solid blue line). B: Pa៮ /PI
with ␴ ⫽ 0, ⌬V̇ ⫽ 0 (black dash-dotted line); ␴ ⫽ 0, ⌬V̇ ⫽ 600 ml/min (solid black line); ␴ ⫽ 1.5, ⌬V̇ ⫽ 0 (red dash-dotted line); ␴ ⫽ 1.5, ⌬V̇ ⫽ 600 ml/min
(solid red line). Dotted green vertical line on the left indicates the position of sulfur hexafluoride, dotted green vertical line on the right indicates the position
of methoxyflurane.
increases in blood but decreases in the gas phase (black The effect of increasing ⌬V̇ on the A-a partial pressure
arrows). This is in line with our previous findings (11). difference is shown in Fig. 5 for several gases in order of
Alveolar-arterial pressure differences. These may be in- decreasing solubility: diethyl ether, sevoflurane, ethylene, and
ferred from the partial pressure-solubility diagram. For exam- sulfur hexafluoride (␭ ⫽ 12, 0.59, 0.14, and 0.0076, respec-
ple, in Fig. 4, A–D, PA៮ /PI and Pa៮ /PI have been drawn with tively). It is assumed that the inspired concentration of each gas
␴ ⫽ 0 (black dash-dotted line) and ␴ ⫽ 1.5 (solid lines; is 1%, which represents an inspired partial pressure of 7.13
blue ⫽ gas, red ⫽ blood) for increasing values of ⌬V̇ (from 0 mmHg. The degree of V̇/Q̇ mismatch, ␴, is shown on the
to 800 ml/min). x-axis. Each curve begins at ␴ ⫽ 0, where there is no V̇/Q̇
When ␴ ⫽ 0, there is no mismatching between ventilation mismatch and the A-a partial pressure difference is zero.
and blood flow; PA៮ /PI and Pa៮ /PI are equal. This is represented Alveolar and arterial partial pressures become equal again
by the single black dash-dotted lines in Fig. 4, A–D regardless when a curve crosses the x-axis, i.e., y ⫽ 0. Parts of a curve
of the value of ⌬V̇. Notice that in each case, the blue line lying below the x-axis indicate a reversal of the A-a partial
representing PA៮ /PI always lies above or coincides with the pressure gradient; i.e., the partial pressure in arterial blood is
black line. However, whereas the red line representing Pa៮ /PI greater than the partial pressure in mixed alveolar gas.
lies below the blue line for PA៮ /PI on the right of the solubility The red arrow in each diagram shows the direction of the
axis, it crosses the blue line as one moves left along the increase in ⌬V̇, which is varied from 0 to 1 l/min in increments
solubility axis. This represents the situation where the partial of 200 ml/min. Ether exhibits a worsening in A-a gradient as
pressure in mixed pulmonary end-capillary blood exceeds that ⌬V̇ is increased, but this is almost imperceptible. In the case of
in mixed alveolar gas. The point of intersection is affected by sevoflurane and ethylene, the A-a partial pressure gradient is
the shrinkage factor and moves to the right as ⌬V̇ is increased. reduced and moves toward zero. With sulfur hexafluoride, the
From this, we can conclude that the solubility at the point of A-a partial pressure difference is less than zero when ⌬V̇ is 400
intersection depends on the magnitude of the net gas volume ml/min or greater, with the partial pressure in pulmonary
uptake. end-capillary blood predicted to be 1 mmHg higher than that in
Fig. 3. A: augmentation-solubility diagram with ␴ ⫽ 1.5. Gas phase, ⌬V̇ ⫽ 200 ml/min (blue dash-dotted line); ⌬V̇ ⫽ 800 ml/min (solid blue line). Blood,
⌬V̇ ⫽ 200 ml/min (red dash-dotted line); ⌬V̇ ⫽ 800 ml/min (solid red line). B: effect of increasing ␴ with ⌬V̇ constant and set at 100 ml/min. Arrows indicate
the direction of increase of ventilation-perfusion mismatch (␴ ⫽ 0, 0.5, 1.0, 1.5, and 2.0).

Fig. 4. Alveolar (blue) and arterial (red) partial pressures as a fraction of the inspired partial pressure (on the y-axis) vs. log10(␭) (on the x-axis) for a range of
␭ values with ␴ ⫽ 1.5, A: ⌬V̇ ⫽ 0, B: ⌬V̇ ⫽ 200, C: ⌬V̇ ⫽ 400, D: ⌬V̇ ⫽ 800 ml/min. Black dash-dotted line in each diagram is associated with no mismatch
of ventilation and perfusion, i.e., ␴ ⫽ 0.
mixed alveolar gas, when ␴ ⫽ 1.2 and ⌬V̇ ⫽ 600 ml/min. It shifted to the right; each curve begins above the R(␭) ⫽ 1 line
can be seen that for any specified value of ⌬V̇, the occurrence and ends below it. At the left end of each curve in Fig. 6B, R(␭)
of A-a partial pressure reversal depends on the solubility of the increases as ␴ increases, whereas at the right end, R(␭) de-
agent and the degree of V̇/Q̇ mismatch. creases as ␴ increases. At the left end of each curve in Fig. 6C,
Uptake-solubility diagrams. Many of the above results are R(␭) increases as ⌬V̇ increases, whereas at the right end, R(␭)
consistent with gas uptake in the presence of ventilation- decreases as ⌬V̇ increases.
perfusion mismatch exceeding that in the absence of ventila- Where R(␭) is greater than 1, uptake of a low inspired
tion-perfusion mismatch, a possibility that is the subject of our concentration of an inert gas is predicted to be greater in a
next investigation. In Fig. 6, we compare the situation that nonhomogenous lung than in the equivalent homogenous lung.
applies when V̇/Q̇ mismatch occurs in the absence of net gas However, only sulfur hexafluoride appears to exhibit this
volume changes with two scenarios in the presence of such property (see blue arrow in Fig. 6B), and only then it is under
changes. extreme conditions, which may not be attainable clinically
Colburn et al. (3) define R(␭) as the ratio of the gas exchange (⌬V̇ ⫽ 1,000 ml/min; ␴ ⫽ 4).
rate of a lung with V̇/Q̇ mismatch over the exchange rate of the
equivalent homogenous lung and describe the shape of the DISCUSSION
graph of R(␭) against log10(␭). Figure 6A shows a series of
such graphs for values of ␴ ranging from 0 to 4 in increments Historically, the properties of the SGE have first been
of 0.5. With very low inspired concentrations of inert gas there predicted from theory and then demonstrated experimentally.
is no net volume change during gas exchange, and the family Thus, the existence of the SGE was first predicted from Eger’s
of curves is symmetric about the vertical line x ⫽ log10(V̇A/Q̇). mathematical model of anesthetic uptake (5, 6). Similarly, the
As ␴ is increased, R(␭) falls as predicted by Colburn et al. (3) disproportionate effect of V̇/Q̇ mismatch on the partial pressure
and is never greater than 1. of gases in blood was first predicted from theory (18 –20). Both
This is the classical result predicted by Colburn et al. (3) propositions were then proven experimentally (7, 9, 17, 22). As
with which we now compare our two scenarios. Figure 6B has been pointed out by Kennedy (10), the theoretical approach
shows the effect of increasing the value of ␴ from zero, with has many advantages. With respect to the present study, a key
the net volume uptake fixed at 1,000 ml/min [body temperature advantage is the ability to control V̇/Q̇ mismatch as an inde-
and pressure dry (BTPD)], whereas Fig. 6C shows the effect of pendent variable, something that is not currently possible in
increasing the value of ⌬V̇ from zero with the degree of experimental situations. Moreover, this is the first time that a
mismatch fixed at ␴ ⫽ 2. In each case, once ⌬V̇ is greater than formal mathematical analysis of the type referred to by West
zero, the curves are no longer symmetric about the vertical line and Wagner (27) has been attempted for the second gas effect.
x ⫽ log10(V̇A/Q̇). The minimum for each curve has been In addition, we have used the powerful technique of Lebesgue-

Fig. 5. Graphs of the A-a partial pressure difference (mmHg) for diethyl ether (A), sevoflurane (B), ethylene (C), and sulfur hexafluoride (D) as a function of
␴ (the degree of V̇/Q̇ mismatch) for ⌬V̇ ⫽ 0, 200, 400, 600, 800, and 1,000 ml/min. In each case, the direction of increase in ⌬V̇ is shown by the red arrow.
Stieltjes integration described by Colburn et al. (3) to perform therefore, it must always be positive. However, this rule clearly
our calculations. Why is this important? The alternative is to cannot apply at the lower end of the solubility scale in Fig. 4,
use a compartmental approach. But how many compartments B–D. Our findings are shown in Fig. 5. With V̇A ⫽ V̇I, the
should be used: 10, 20, 50, 100? The more compartments, the A-a partial pressure difference always increases as the degree
greater the accuracy, but this requires more numerical calcu- of V̇/Q̇ mismatch increases, but less so as the solubility of the
lations and hence, more computer time. By transforming the gas is decreased. As the difference between V̇A and V̇I is
problem into one with an algebraic solution, it can be dealt increased with the less soluble gases shown in Fig. 5 B–D, the
with expeditiously using the inbuilt functions of a modern A-a difference decreases. In the case of sulfur hexafluoride, the
problem-solving program like MATLAB. least soluble of the gases, the pressure gradient is reversed,
Our study provides several insights into the SGE. Perhaps with the partial pressure in blood predicted to be higher than
the most interesting finding is the effect on the predicted A-a that in the gas phase, although the movement of the sulfur
partial pressure difference. According to West et al. (26), hexafluoride is still from gas phase to blood. One would
during gas uptake, the alveolar component of the A-a differ- normally expect the partial pressure of a gas to be lower at all
ence is equal to ␭ over V̇A/Q̇ times the arterial component, and points downstream during gas uptake, for example, as is
Fig. 6. Graph of R(␭) vs log10(␭) for a range of ␭ values. A: in the absence of net volume change. Red arrows indicate the direction of increase of ␴ from 0 to
4 in increments of 0.5. B: with a net volume change of 1,000 ml/min [body temperature and pressure dry (BTPD)], so that ⌬V̇/V̇A ⫽ 0.25. Red arrows are same
as in A. Blue arrow indicates the point where the uptake ratio of sulfur hexafluoride exceeds 1. C: with the degree of V̇/Q̇ mismatch fixed at ␴ ⫽ 2, the net volume
change has been increased in the direction of the red arrows from 0 to 1,000 ml/min (BTPD) in increments of 200 ml/min. In each diagram, the solid vertical
green line is drawn at x ⫽ log10(0.8). The position of sulfur hexafluoride is shown by the dotted green vertical line on the left, and that of methoxyflurane is
shown by the dotted green vertical line on the right.

commonly described in the transfer of oxygen from inspired air “unsafe” FIO2. In the region above the dark blue surface and
to the mitochondrion, the so-called “oxygen cascade” (1). below the yellow FIN O ⫽ 0.7 plane (green arrow), it is possible
As is apparent from the comments in RESULTS in relation to 2
for nitrous oxide uptake to generate a reversal of the alveolar-
Figs. 4 and 5, whether this phenomenon occurs in a clinical arterial partial pressure gradient for the second gas.
setting will depend on the value of ␭, ␴, and ⌬V̇/V̇A, the factors
During anesthesia, ␴ generally lies between 0.75 and 1.5.
controlling augmentation in Eq. 9. Although an increase in the
With ␴ ⫽ 0.75, the highest solubility for which A-a reversal
degree of V̇/Q̇ mismatch favors augmentation through its effect
can occur is ⬃0.35 so that none of the volatile agents in
on the term k in Eq. 9, it impedes nitrous oxide uptake, and it
common use can be expected to demonstrate A-a partial
can therefore reduce augmentation through its effect on the
pressure reversal during uptake. However at least six other
“shrinkage” term ⌬V̇/V̇A. Increasing the inspired nitrous oxide
gases may demonstrate the effect: xenon (␭ ⫽ 0.2), ethylene
concentration can compensate for this but is limited by the
(␭ ⫽ 0.14), ethane (␭ ⫽ 0.092), methane (␭ ⫽ 0.038), nitrogen
maximum concentration compatible with safe anesthesia.
(␭ ⫽ 0.014), and sulfur hexafluoride (␭ ⫽ 0.0076). With
These competing effects are shown in Fig. 7. Each point on the
␴ ⫽ 1.5, the highest solubility for which A-a reversal can occur
blue colored curved surface in Fig. 7 represents a point at
which the alveolar and arterial partial pressures are equal. This is ⬃0.15 so that only ethane, methane, nitrogen, and sulfur
surface is shown as a function of the gas solubility (x-axis) and hexafluoride can be expected to demonstrate the phenomenon.
the degree of V̇/Q̇ mismatch (y-axis). Shown on the z-axis is So far, we have only considered a second gas with a mixed
venous partial pressure of zero undergoing net uptake in the
the FIN O corresponding to each point on the surface. The
2 presence of volume “shrinkage” produced by nitrous oxide
surface slopes upward and backward from the bottom right to uptake. A slightly different scenario was investigated by Can-
the top left of Fig. 7. A yellow horizontal plane is drawn for field and Rahn (2), who studied alveolar-arterial nitrogen
FIN O ⫽ 0.7, the highest inspired concentration that can be used partial pressure differences in anesthetized dogs being venti-
2
safely in most healthy subjects without producing hypoxia lated with air. Here, the difference between inspired and
(16). In the region beneath the blue surface (red arrow), the expired alveolar ventilation is smaller than during nitrous oxide
inspired nitrous oxide concentration is too low to produce the anesthesia. It is produced by the discrepancy between oxygen
required volume change. The region above the yellow plane on uptake and carbon dioxide output, but there is no net exchange
which FIN O ⫽ 0.7 (Fig. 7, black arrow) is associated with an of nitrogen. They showed that the mean arterial pressure of
2
Fig. 7. Three-dimensional surface showing all points for which the alveolar and arterial partial pressures of the 2nd gas are equal, shown as a function of the
inspired nitrous oxide concentration necessary to produce the required volume change for the specified values of ␭ and ␴. The surface is colored blue and slopes
upward and backward from the bottom right (dark blue) to the top left (light blue). On the yellow horizontal plane, FIN O ⫽ 0.7. This plane intersects the blue
2
surface as shown. Red arrow points to the region below the blue surface in which the FIN O is too low to produce the required volume change. Green arrow points
2
to the region between the dark blue curved surface and the yellow horizontal plane compatible with the phenomenon of A-a partial pressure reversal. Black arrow
points to the region above the yellow plane associated with an “unsafe” FIO . When ␴ ⫽ 0, PA៮ ⫽ Pa៮ for all values of FIN O, so the graph has been offset slightly
2 2
from ␴ ⫽ 0 to facilitate discussion.

Fig. 8. A: alveolar-arterial partial pressure difference for nitrogen as a function of ␴, the degree of ventilation-perfusion mismatch, for ⌬V̇ ⫽ 0, 25, 50, 75, and
100 ml/min. The direction of increase in ⌬V̇ is shown by the red arrow. The dotted horizontal green line indicates an arterial-alveolar difference of 16 mmHg.
B: A-a partial pressure difference for nitrogen as a function of ␴, the degree of ventilation-perfusion mismatch, for ⌬V̇ ⫽ 0, ⫺25, ⫺50, ⫺75, and ⫺100 ml/min.
The direction of decrease in ⌬V̇ is shown by the blue arrow.
nitrogen was greater than that of mean alveolar gas and Uptake of the second gas may occur against an overall alveo-
ascribed it to the variation in V̇/Q̇ ratios throughout the lung. lar-arterial partial pressure gradient if the degree of V̇/Q̇
The situation may be mimicked in our model by setting Pv៮ mismatch and the rate of nitrous oxide uptake are sufficiently
equal to Pa៮ for the inert gas involved (see APPENDIX 2). The large and can occur with highly insoluble gases such as sulfur
result is shown in Fig. 8A for values of ⌬V̇ ⫽ 0, 25, 50, 75, and hexafluoride but is less likely with agents such as desflurane
100 ml/min. and sevoflurane in common use today. We have also shown
This may be compared with the graphs in Fig. 5. The red that it is possible for V̇/Q̇ mismatch to increase second gas
arrow shows the direction of increase in ⌬V̇. This is associated uptake compared with the equivalent homogenous lung.
with an increase in the size of the a-A partial pressure gradient.
Canfield and Rahn (2) obtained an average gradient of 16 APPENDIX 1
mmHg, which is indicated by the horizontal green dotted line in
Fig. 8A. The curve for ⌬V̇ ⫽ 50 ml/min intersects this line with Derivation of Lebesgue-Stieltjes integrals for second gas partial
␴ approximately equal to 1.8. This is close to the upper limit of pressure-solubility diagrams. In a lung model consisting of N parallel
␴ values known to occur in humans during anesthesia. The compartments, each having a different ventilation/perfusion ratio, the
partial pressure of the mixed expired alveolar gas PA៮ is the ventila-
authors’ statement, “Whenever differences in ventilation-per-
tion-weighted mean of the individual compartmental partial pressures:
fusion ratio exist among alveoli of the lung . . . the mean
N
arterial inert gas pressure is greater than that of the mean
alveolar gas” is borne out in Fig. 8A with one important 兺
j⫽1
V̇A PA j j
PA៮ ⫽ . (10)
proviso, namely that the ventilation-perfusion differences must N
occur in the presence of a reduction in the system volume; i.e., 兺

j⫽1
V̇A j
⌬V̇ must be greater than zero. Thus, the line for ⌬V̇ ⫽ 0 in Fig.
8A is associated with no difference between mean alveolar and The partial pressure of the mixed pulmonary end-capillary blood Pa៮
is the perfusion-weighted mean of the individual compartmental
mean arterial partial pressure for nitrogen. Furthermore, if the partial pressures:
volume change occurs in the reverse direction, i.e., ⌬V̇ ⬍ 0,
N
the direction of the alveolar-arterial nitrogen partial pressure
gradient is reversed. This is illustrated in Fig. 8B. However, an 兺
j⫽1
Q̇ PAj j
P៮a ⫽ . (11)
internal circulation of nitrogen will always occur in the pres- N
ence of ventilation-perfusion mismatch whatever the value of 兺

j⫽1
Q̇ j
⌬V̇. This example further demonstrates the importance of

These expressions were used in deriving Eqs. 7 and 8. For a further
taking net gas volume changes into account when considering treatment of the SGE, we now introduce the subscript “SG” to
gas exchange. distinguish the second gas from the first gas, nitrous oxide, which is
left, for simplicity, with no subscript.
CONCLUSION From Eqs. 10 and 11, we obtain the following:
We have investigated the effect of net gas volume loss such
as that seen during nitrous oxide anesthesia on low concentra-
tions of second gases present. It was shown that the classical
FjSG
FISG
⫽
V̇j ⁄ Q̇j
冉 ␭ ⫹ V̇j ⁄ Q̇j
␭SG ⫹ V̇j ⁄ Q̇j ␭共1 ⫺ FI兲 ⫹ V̇j ⁄ Q̇j
冊 . (12)
partial pressure-solubility diagram of West et al. (26) must be Here, the symbols Q̇j and V̇j refer to the blood flow and expired
modified. At the lower end of the solubility range, the partial alveolar ventilation in the jth compartment, ␭ and FI to the solubility
pressure-solubility curve of the second gas in arterial blood is and inspired concentration of nitrous oxide, and ␭SG to the solubility
displaced upward, resulting in a disproportionate increase in of the second gas in blood. Setting ␭= ⫽ ␭(1 ⫺ FI), and expressing Eq.
the second gas effect in blood compared with alveolar gas. 12 using partial fractions,

共␭ ⫺ ␭SG兲 P៮ASG N PjSG N

FjSG
⫽
V̇j ⁄ Q̇j ⫽ 兺共V̇j ⁄ V̇A兲 ⫽ 兺共V̇j ⁄ V̇A兲
共 SG兲共 ␭SG ⫹ V̇j ⁄ Q̇j 兲
再冉冊冎
FISG ␭⬘ ⫺ ␭ PISG j⫽1 PISG j⫽1
␭ ⫹ V̇j ⁄ Q̇j ␭SG共P V៮ ⁄ PI兲SG

共␭ ⫺ ␭⬘兲共V̇j ⁄ Q̇j兲
V̇j ⁄ Q̇j
⫹ , (23)
⫺ . (13) ␭SG ⫹ V̇j ⁄ Q̇j ␭共1 ⫺ FI兲 ⫹ V̇j ⁄ Q̇j ␭SG ⫹ V̇j ⁄ Q̇j
共␭⬘ ⫺ ␭SG兲共␭⬘ ⫹ V̇j ⁄ Q̇j兲
and for the mean pulmonary end-capillary blood, we have
Applying Eqs. 10 and 11 with V̇A and Q̇t as symbols for the total
Pa៮SG N PjSG N
共Q̇j ⁄ Q̇t兲 ⫽ 兺共Q̇j ⁄ Q̇t兲
alveolar ventilation and total pulmonary blood flow respectively, we
obtain ⫽ 兺
再冉冊冎
PISG j⫽1 PISG j⫽1
៮ SG
PA 共V̇A ⁄ Q̇t兲 V̇j ⁄ Q̇j ␭ ⫹ V̇j ⁄ Q̇j ␭SG共P V៮ ⁄ PI兲SG
⫽ 兵共␭ ⫺ ␭SG兲I1 ⫺ 共␭ ⫺ ␭⬘兲I2其 , (14) ⫹ .
PISG 共␭⬘ ⫺ ␭SG兲兹2␲ ␭SG ⫹ V̇j ⁄ Q̇j ␭共1 ⫺ FI兲 ⫹ V̇j ⁄ Q̇j ␭SG ⫹ V̇j ⁄ Q̇j
Pa៮SG 共V̇A ⁄ Q̇t兲 (24)
⫽⫺ 兵共␭ ⫺ ␭SG兲I3 ⫺ 共␭ ⫺ ␭⬘兲I4其 , (15) Defining I5 as follows:
PISG 共␭⬘ ⫺ ␭SG兲兹2␲
⬁
兰␭
⬁ dx
兰␭
dx I5 ⫽ , (25)
I1 ⫽
⫺⬁ SGe
共x 2⫹2␴x⫺␴2 ⁄ 2
共
兲 ⫹ V̇A ⁄ Q̇t ex2⁄ 2 兲
, (16)
⫺⬁ SG e 共 x ⫺ ␴兲 ⁄ 2 ⫹
2
共V̇A ⁄ Q̇t兲e共x ⫺4␴x⫹2␴ 兲 ⁄ 2
2 2
⬁ Equations 23 and 24 may be converted to the following equations

兰 ␭⬘e共
dx
I2 ⫽ , (17) containing Lebesgue-Stieltjes integrals:
⫺⬁
x2⫹2␴x⫺␴2兲 ⁄ 2
共
⫹ V̇A ⁄ Q̇t ex 兲 2⁄2
⬁ P A៮ SG 共V̇A ⁄ Q̇t兲
⫽ 兵共␭ ⫺ ␭SG兲I1 ⫺ 共␭ ⫺ ␭⬘兲I2其
兰␭
dx
I3 ⫽ , (18) PISG 共␭⬘ ⫺ ␭SG兲兹2␲
⫺⬁ SGe
x2⁄ 2
共
⫹ V̇A ⁄ Q̇t e 兲共 x ⫺ ␴兲 2 ⁄ 2
␭SG共P V៮ ⁄ PI兲SG
and ⫹ I3 . (26)
⬁
兹2␲
兰 ␭⬘e
dx
I4 ⫽ . (19) and
⫺⬁
x2⁄ 2
共
⫹ V̇A ⁄ Q̇t e共x ⫺ ␴兲兲 2⁄ 2
Pa៮SG 共V̇A ⁄ Q̇t兲

The following equation is then solved for ␭= in MATLAB: ⫽ 兵共␭ ⫺ ␭SG兲I3 ⫺ 共␭ ⫺ ␭⬘兲I4其
PISG 共␭⬘ ⫺ ␭SG兲兹2␲
P៮ASG Pa៮SG ␭SG共P V៮ ⁄ P 兲SG
⫺ ⫽ 0, (20) I
PISG PISG ⫹ I5 . (27)
兹2␲
and the result is converted to a value for FIN O by inverting the
2 Because P៮VSG ⫽ Pa៮SG we can eliminate the term containing Pv៮ from
relationship between ␭= and FIN O:
2 Eq. 27, and substitute for (Pv៮ /PI)SG in Eq. 26.
␭⬘
F IN O ⫽ 1 ⫺ . (21) Pa៮SG 共V̇A ⁄ Q̇t兲
␭ ⫽ 兵共␭ ⫺ ␭SG兲I3 ⫺ 共␭ ⫺ ␭⬘兲I4其.
共␭⬘ ⫺ ␭SG兲共兹2␲ ⫺ ␭SGI5兲
2
PISG
(28)
APPENDIX 2
Note that if ␴ ⫽ 0, I1 – I5 become
Derivation of Lebesgue-Stieltjes integrals with second gas pres-
ent-in mixed venous blood. In the above equations, the mixed venous
I1 ⫽ I3 ⫽ I5 ⫽
兹2␲ , (29)
partial pressure of each gas has been set to zero. Canfield and Rahn (2)
describe the existence of a circulation of nitrogen in the lung due to 共␭SG ⫹ V̇A ⁄ Q̇t兲
V̇/Q̇ differences in different parts of the lung and conclude that the and
partial pressure of N2 in blood must be greater than that in the gas
phase because of these differences, although there is no net uptake of
the gas in the steady-state, breathing air. We can incorporate this I2 ⫽ I4 ⫽
兹2␲ . (30)
scenario in our model by including the mixed venous partial pressure 共␭⬘ ⫹ V̇A ⁄ Q̇t兲
of N2 in the mass balance and setting it equal to the arterial partial
So, Eq. 28 becomes
pressure. Pv៮ is then an unknown to be determined for N2. It is still
necessary to determine PI for N2O uptake to produce the desired value
of ⌬V̇. For the second gas, we now have, after equilibration in the jth
Pa៮SG
⫽
共␭ ⫹ V̇A ⁄ Q̇t兲 , (31)
compartment, PISG 共␭⬘ ⫹ V̇A ⁄ Q̇t兲
冉冊
PASG
PISG j
⫽
V̇j ⁄ Q̇j
冉
␭SG ⫹ V̇j ⁄ Q̇j ␭共1 ⫺ FI兲 ⫹ V̇j ⁄ Q̇j
␭ ⫹ V̇j ⁄ Q̇j
冊 and Eq. 26 produces the same result so that if FIN O ⬎ 0, ␭= ⬍ ␭,
2
alveolar and arterial pressures are equal and both greater than the
inspired partial pressure. If ␭= ⫽ ␭, alveolar, arterial, and inspired
␭SG共P V ៮ ⁄ PI兲SG partial pressures are all equal.
⫹ . (22)
␭SG ⫹ V̇j ⁄ Q̇j The situation that applies when the net volume movement is from
blood to alveolar gas is shown in Fig. 8B. Here, we have set PI to zero
So, for the mean alveolar partial pressure we now have for N2O and determined the value of Pv៮ associated with the specified

value of ⌬V̇. With ␭ ⫽ ␭N2O as before and letting ␭* ⫽ ␭共1 ⫺ integration purposes. We have not gone down this path, as the other
FV៮ N O兲, this gives rise to the following set of equations for the second conclusions drawn from Fig. 8 regarding the direction of the alveolar-
2
gas, nitrogen: arterial partial pressure gradient still apply, as they are due to the
assumption that Pv៮ ⫽ Pa៮ for nitrogen.
␭ FV៮ N OV̇A
⌬V̇ ⫽ ⫺ 2
Iv , (32) DISCLOSURES
兹2␲ No conflicts of interest, financial or otherwise, are declared by the authors.
⬁
兰 ␭e
dx
Iv ⫽ AUTHOR CONTRIBUTIONS
x2⁄ 2 2⁄ 2
⫺⬁ ⫹ (V̇A ⁄ Q̇t)e(x ⫺ ␴) B.K. conception and design of research; B.K. performed experiments; B.K.,
R.K.D., and P.J.P. analyzed data; B.K., R.K.D., and P.J.P. interpreted results
FV៮ N O ⫽ ⫺
兹2␲⌬V̇ , (33)
of experiments; B.K. prepared figures; B.K. drafted manuscript; B.K., R.K.D.,
and P.J.P. edited and revised manuscript; B.K., R.K.D., and P.J.P. approved
2
␭V̇AIv final version of manuscript.
៮ SG
PA (V̇A ⁄ Q̇t)
⫽ 兵共␭* ⫺ ␭兲H1 ⫺ 共␭* ⫺ ␭SG兲H2其 REFERENCES
PISG 共␭SG ⫺ ␭兲兹2␲ 1. Biro GP. From the atmosphere to the mitochondrion: the oxygen cascade.
␭SG共P V
៮ ⁄ PI兲SG In: Hemoglobin-Based Oxygen Carriers as Red Cell Substitutes and
⫹ H3 , (34) Oxygen Therapeutics (edited by Kim H and Greenburg A). Berlin, Heidel-
兹2␲ berg: Springer, 2013, p. 27–53.
2. Canfield RE, Rahn H. Arterial-alveolar N2 gas pressure differences due
Pa៮SG (V̇A ⁄ Q̇t)
⫽ 兵共␭* ⫺ ␭兲H4 ⫺ 共␭* ⫺ ␭SG兲H5其 to ventilation-perfusion variations. J Appl Physiol 10: 165–172, 1957.
PISG 共 SG 兲兹
␭ ⫺ ␭ 2␲ doi:10.1152/jappl.1957.10.2.165.
3. Colburn WE Jr, Evans JW, West JB. Analysis of effect of the solubility
␭SG共PV
៮ ⁄ PI兲SG on gas exchange in nonhomogenous lungs. J Appl Physiol 37: 547–551,
⫹ H6 , (35)
兹2␲ 1974. doi:10.1152/jappl.1974.37.4.547.
4. Dueck R, Young I, Clausen J, Wagner PD. Altered distribution of
⬁
pulmonary ventilation and blood flow following induction of inhalation
兰 ␭e
dx
H1 ⫽ , (36) anesthesia. Anesthesiology 52: 126 –130, 1980. doi:10.1097/00000542-
(x2⫹2␴x⫺␴2) ⁄ 2 2⁄ 2
⫺⬁ ⫹ (V̇A ⁄ Q̇t)ex 198002000-00004.
⬁ 5. Eger EI. Applications of a mathematical model of gas uptake. In: Uptake
兰␭
dx and Distribution of Anesthetic Agents (edited by Papper EM and Kitz RJ).
H2 ⫽ , (37) New York: McGraw-Hill, 1963, p. 88 –103.
(x2⫹2␴x⫺␴2) ⁄ 2 2⁄ 2
⫺⬁ SGe ⫹ (V̇A ⁄ Q̇t)ex 6. Eger EI. A mathematical model of uptake and distribution. In: Uptake and
⬁ Distribution of Anesthetic Agents (edited by Papper EM and Kitz RJ). New
兰␭
dx
H3 ⫽ , (38) York: McGraw-Hill, 1963, p. 72–87.
x2⁄ 2 2⁄ 2
⫺⬁ SGe ⫹ (V̇A ⁄ Q̇t)e(x ⫺ ␴) 7. Epstein RM, Rackow H, Salanitre E, Wolf GL. Influence of the concen-
tration effect on the uptake of anesthetic mixtures: the second gas effect.
⬁
兰 ␭e
dx Anesthesiology 25: 364–371, 1964. doi:10.1097/00000542-196405000-00015.
H4 ⫽ , (39) 8. Evans JW, Wagner PD, West JB. Conditions for reduction of pulmonary
x2⁄ 2 2⁄ 2
⫺⬁ ⫹ (V̇A ⁄ Q̇t)e(x ⫺ ␴) gas transfer by ventilation-perfusion inequality. J Appl Physiol 36: 533–
⬁ 537, 1974. doi:10.1152/jappl.1974.36.5.533.
兰␭
dx 9. Hendrickx JF, Carette R, Lemmens HJM, De Wolf AM. Large volume
H5 ⫽ , (40)
x2⁄ 2 2⁄ 2
⫺⬁ SGe ⫹ (V̇A ⁄ Q̇t)e(x ⫺ ␴) N2O uptake alone does not explain the second gas effect of N2O on
sevoflurane during constant inspired ventilation. Br J Anaesth 96: 391–
and 395, 2006. doi:10.1093/bja/ael008.
10. Kennedy RR. A second look at the second gas effect. Anesthesiology 128:
⬁
兰␭
dx 1053–1054, 2018. doi:10.1097/ALN.0000000000002206.
H6 ⫽ . (41) 11. Korman B, Dash RK, Peyton PJ. Can mathematical modeling explain
(x ⫺ ␴)2⁄ 2 2⫺4␴x⫹2␴2) ⁄ 2
⫺⬁ SGe ⫹ (V̇A ⁄ Q̇t)e(x the measured magnitude of the second gas effect? Anesthesiology 128:
1075–1083, 2018. doi:10.1097/ALN.0000000000002131.
Because PV៮SG ⫽ Pa៮SG, we can proceed as before and eliminate the 12. Korman B, Mapleson WW. Concentration and second gas effects: can
term containing Pv៮ from Eq. 35: the accepted explanation be improved? Br J Anaesth 78: 618 –625, 1997.
doi:10.1093/bja/78.5.618.
Pa៮SG 共V̇A ⁄ Q̇t兲 13. Lundh R, Hedenstierna G. Ventilation-perfusion relationships during
⫽ 兵共␭* ⫺ ␭兲H4 ⫺ 共␭* ⫺ ␭SG兲H5其.
共␭SG ⫺ ␭兲共兹2␲ ⫺ ␭SGH6兲
PISG halothane anaesthesia and mechanical ventilation. Effects of varying
inspired oxygen concentration. Acta Anaesthesiol Scand 28: 191–198,
(42) 1984. doi:10.1111/j.1399-6576.1984.tb02039.x.
Although we have multiplied Eqs. 26, 27, 34, and 35 by 563, the 14. Lundh R, Hedenstierna G. Ventilation-perfusion relationships during
anaesthesia and abdominal surgery. Acta Anaesthesiol Scand 27: 167–173,
partial pressure of nitrogen in dry air, to produce the curves in Fig. 8,
1983. doi:10.1111/j.1399-6576.1983.tb01929.x.
the quantitative results are only approximate since our analysis is 15. Hedenstierna G, Lundh R, Johansson H. Alveolar stability during
based on a second gas present in infinitesimal concentrations. This is anaesthesia for reconstructive vascular surgery in the leg. Acta Anaesthe-
equivalent to the simplifying assumption that V̇AIFIN ⫽ V̇AFAN in siol Scand 27: 26 –34, 1983. doi:10.1111/j.1399-6576.1983.tb01900.x.
2 2
each gas-exchanging compartment, an assumption made initially by 16. Magnusson L, Spahn DR. New concepts of atelectasis during general
Rahn and Fenn (23) to enable V̇/Q̇ lines to be drawn on the oxygen- anaesthesia. Br J Anaesth 91: 61–72, 2003. doi:10.1093/bja/aeg085.
17. Peyton PJ, Horriat M, Robinson GJ, Pierce R, Thompson BR. Magnitude
carbon dioxide diagram. Once this no longer applies, Eq. 2 must be of the second gas effect on arterial sevoflurane partial pressure. Anesthesiol-
modified to include the second gas, giving rise to two equations that ogy 108: 381–387, 2008. doi:10.1097/ALN.0b013e318164caf3.
must be solved simultaneously to yield the mean alveolar and mixed 18. Peyton PJ, Robinson GJ, Thompson B. Effect of ventilation-perfusion
pulmonary end-capillary partial pressures of each gas. These compli- inhomogeneity and N2O on oxygenation: physiological modeling of gas ex-
cated equations must be broken down into a series of expressions for change. J Appl Physiol (1985) 91: 17–25, 2001. doi:10.1152/jappl.2001.91.1.17.

19. Peyton PJ, Robinson GJ, Thompson B. Ventilation-perfusion inhomoge- 23. Rahn H, Fenn WO. A Graphical Analysis of the Respiratory Gas Exchange.
neity increases gas uptake in anesthesia: computer modeling of gas exchange. Washington, DC: American Physiological Society, 1955, p. 39.
J Appl Physiol (1985) 91: 10 –16, 2001. doi:10.1152/jappl.2001.91.1.10. 24. Rehder K, Knopp TJ, Sessler AD, Didier EP. Ventilation-perfusion
20. Peyton PJ, Robinson GJ, Thompson B. Ventilation-perfusion inhomo- relationship in young healthy awake and anesthetized-paralyzed man. J
geneity increases gas uptake: theoretical modeling of gas exchange. J Appl Appl Physiol 47: 745–753, 1979. doi:10.1152/jappl.1979.47.4.745.
Physiol (1985) 91: 3–9, 2001. doi:10.1152/jappl.2001.91.1.3. 25. Stoelting RK, Eger EI II. An additional explanation for the second gas
21. Peyton PJ, Stuart-Andrews C, Deo K, Strahan F, Robinson GJ, effect: a concentrating effect. Anesthesiology 30: 273–277, 1969. doi:10.
Thompson BR, Pierce R. Persisting concentrating and second gas effects 1097/00000542-196903000-00007.
on oxygenation during N2O anaesthesia. Anaesthesia 61: 322–329, 2006. 26. West JB, Wagner PD, Derks CM. Gas exchange in distributions of V̇A/Q̇
doi:10.1111/j.1365-2044.2006.04579.x. ratios: partial pressure-solubility diagram. J Appl Physiol 37: 533–540,
22. Peyton PJ, Fortuin M, Robinson GJB, Stuart-Andrews C, Pierce R, 1974. doi:10.1152/jappl.1974.37.4.533.
Thompson BR. The rate of alveolar-capillary uptake of sevoflurane and 27. West JB, Wagner PD. Pulmonary gas exchange. In: Bioengineering
nitrous oxide following anaesthetic induction. Anaesthesia 63: 358 –363, Aspects of the Lung (edited by West JB). New York: Marcel Dekker, 1977,
2008. doi:10.1111/j.1365-2044.2007.05355.x. p. 378 –382.

J Appl Physiol 128: 1587–1593, 2020.
First published March 19, 2020; doi:10.1152/japplphysiol.00049.2020.
RESEARCH ARTICLE
Elucidating the roles of solubility and ventilation-perfusion mismatch in the

second gas effect using a two-step model of gas exchange
Ben Korman,1 Ranjan K. Dash,2 and Philip J. Peyton3
1
Department of Anaesthesia and Pain Medicine, Royal Perth Hospital, Perth, Western Australia, Australia; 2Departments of
Biomedical Engineering and Physiology, Medical College of Wisconsin, Wisconsin; and 3Anaesthesia, Perioperative, and
Pain Medicine Unit, Melbourne Medical School, University of Melbourne, Department of Anaesthesia, Austin Health,
Heidelberg, Victoria, Australia
Submitted 21 January 2020; accepted in final form 16 March 2020
Korman B, Dash RK, Peyton PJ. Elucidating the roles of solu- of gas, of the order of 0.5 to 1 L/min, from lungs to blood
bility and ventilation-perfusion mismatch in the second gas effect during the early stages of anesthesia (12, 23). Thereafter, the
using a two-step model of gas exchange. J Appl Physiol 128: 1587– rate of N2O uptake falls progressively toward zero. Alveolar-
1593, 2020. First published March 19, 2020; doi:10.1152/jappl- capillary uptake of the FG increases the alveolar concentrations
physiol.00049.2020.—The second gas effect occurs when high in-
spired concentrations of a first gas, usually nitrous oxide, enhance the
of other gases present, accelerating their uptake. These other
uptake of other gases administered simultaneously. The second gas gases include the commonly used volatile anesthetic agents
effect is greater in blood than in the gas phase, persists well into the desflurane and sevoflurane (␭ ⫽ 0.42 and 0.65) and are re-
period of nitrous oxide maintenance anesthesia, increases as the ferred to individually and collectively as second gases (SGs).
degree of ventilation-perfusion mismatch increases, and is most pro- They are commonly added because, used alone, N2O is not
nounced with the low soluble agents in current use. Yet, how low gas sufficiently potent to produce and maintain anesthesia. We use
solubility and increased ventilation-perfusion mismatch can combine the symbol ␭ for solubility in this paper. By solubility, we
to improve gas transfer remains unclear, which is the focus of the mean the blood/gas partition coefficient at 37°C.
present study. Specifically, we have used a two-step model of gas For decades, the most-widely accepted explanation of the
exchange to separate the effect of gas volume contraction, which SGE appealed to the contraction in gas volume associated with
accompanies the first gas uptake, from other factors. Step 1 involves
the uptake of the second gas at constant volume. Contraction of gas N2O uptake as the factor responsible (12). The SGE was
volume takes place in step 2 and is most effective in transferring measured in expired gas and generally assumed to be of similar
further amounts of gas to blood if the volume of second gas exposed magnitude in blood. Since 2001, it has been shown that the
to the contraction is maximized, i.e., if the loss of second gas in step effect is not only greater in blood but also continues to be
1 is minimized. Minimization depends on having a gas with a low significant for sevoflurane and desflurane when N2O uptake
solubility in blood and increases as the degree of ventilation-perfusion falls to relatively low steady-state levels and that this is due to
mismatch increases. The effectiveness of the contraction also requires the mismatch between ventilation and blood flow (V̇/Q̇ mis-
a favorable alignment with the retained second gas. Alignment de- match) that occurs routinely following the induction of anes-
pends on the solubility of both gases and the degree of ventilation- thesia (8, 10, 17–21, 23). Other manifestations of increased SG
perfusion mismatch. The model is fully consistent with classical
uptake have been described due to increased V̇/Q̇ mismatch but
concepts of gas exchange.
also only for gases with a low solubility in blood (10, 11).
NEW & NOTEWORTHY Gas exchange in the lung can always be Exactly how a low solubility combines with a high degree of
represented as the sum of two components: gas exchange at constant V̇/Q̇ mismatch to increase SG uptake remains unexplained.
volume followed by gas exchange on volume correction. Using this According to basic concepts of respiratory physiology, the
sequence to study the second gas effect, low gas solubility and
effect on gas transfer should be the opposite (16, 27, 28).
increased ventilation-perfusion mismatch are shown to act together to
enhance second gas uptake. While appearing to contravene classical These counterintuitive findings have prompted us to inves-
concepts of gas exchange, a detailed theoretical analysis shows it is tigate more closely the roles of the first and second gases and
fully consistent with these concepts. how these gases with different solubilities interact in the
presence of V̇/Q̇ mismatch to produce the SGE.
anesthetic uptake; mathematical modeling; second gas effect; venti-
lation-perfusion mismatch
METHODS
To study the SGE, it is necessary to identify the component of SG

INTRODUCTION uptake associated with the contraction in gas volume due to N2O
absorption. To do this, we first describe gas exchange in a homoge-
The second gas effect (SGE) occurs when a soluble first gas nous lung under steady-state conditions. We then extend this descrip-
(FG), such as nitrous oxide (N2O), is delivered in high inspired tion to a nonhomogenous lung with a log normal distribution of V̇/Q̇
concentrations during anesthesia (6, 26). The high inspired ratios. Finally, the effect of volume contraction is investigated using
concentrations are associated with the transfer of large volumes a two-step model of alveolar gas exchange in which volume contrac-
tion occurs in step 2.
Gas exchange in a homogenous lung. Consider a first gas with a
Correspondence: B. Korman (e-mail: ben@korman.com.au). linear dissociation curve characterized by a blood-gas solubility co-
http://www.jap.org 8750-7587/20 Copyright © 2020 the American Physiological Society 1587

1588 ANALYSIS OF THE SECOND GAS EFFECT USING A TWO-STEP MODEL
efficient ␭, being exchanged under steady-state conditions. If such a Eq. 3 but the FI term may be ignored. In addition, the second term of
gas is delivered to the alveoli in the inspired gas mixture at a rate V̇IFI Eq. 5 may also be applied to determine the uptake of SG. These
and removed at a rate V̇AFA in a homogenous lung of given V̇A/Q̇t, relationships are discussed further in APPENDIX A.
where V̇I, V̇A, and Q̇t are the total inspired alveolar ventilation, total Description of gas uptake as a two-step process in a single lung
expired alveolar ventilation, and total pulmonary blood flow, respec- unit. The uptake of the FG in a single gas-exchanging compartment is
tively, then the uptake of the gas is V̇IFI ⫺ V̇AFA, where FI and FA are given by either side of Eq. 1. More generally, the steady-state
the fractional concentrations of FG in the dry portion of the inspired exchange V̇G, of any gas G, is given by the left side of Eq. 1. We may
alveolar gas mixture and expired alveolar gas. This must be equal to therefore write:
the gas uptake by blood, which is given by ␭Q̇t共Fa ⫺ Fv៮兲, where Fa V̇G ⫽ V̇IFI ⫺ V̇AFA (6)
and Fv៮ are given by Pa ⁄ 共PB ⫺ PH2O兲and Pv៮ ⁄ 共PB ⫺ PH2O兲, respectively,
with Pa and Pv៮ equal to the partial pressures of the gas in arterial and Consider now this process as taking place in two steps, the first step
mixed venous blood and PB and PH2O the barometric pressure and at constant volume V̇I and the second step during the volume correc-
saturated vapor pressure of water at 37°C, respectively, so that: tion from V̇I to V̇A. This assumption gives rise to the following pair
of equations:
V̇IFI ⫺ V̇AFA ⫽ ␭Q̇t(Fa ⫺ Fv៮) (1)
V̇G(01) ⫽ V̇IFI ⫺ V̇IF⬘ (7)
For the diluent gas, which is treated as if it were completely insoluble,
we have: V̇G(12) ⫽ V̇IF⬘ ⫺ V̇AFA (8)
V̇I共1 ⫺ FI兲 ⫽ V̇A共1 ⫺ FA兲 (2) In Eq. 7, FI and F= are the fractional concentrations in dry gas at the
Assuming complete equilibration between gas and blood phases so start and end of step 1, and V̇G共01兲 is the uptake (or output) of the gas
in this step. Similarly, in Eq. 8, F= and FA are the fractional concen-
that Pa ⫽ PA, V̇I may be eliminated from Eq. 1, which gives rise to the
following general equation for the alveolar and arterial partial pres- trations in dry gas at the start and end of step 2, and V̇G共12兲 is the
sures of the FG: uptake (or output) of the gas in this step. It follows from the addition
of Eqs. 7 and 8 by comparison with Eq. 6 that V̇G ⫽ V̇G共01兲 ⫹ V̇G
共V̇A ⁄ Q̇t兲PI ⫹ ␭共1 ⫺ FI兲Pv៮ 共12兲 and may therefore also be referred to as V̇G共02兲.
Pa ⫽ PA ⫽ (3)
共V̇A ⁄ Q̇t兲 ⫹ ␭共1 ⫺ FI兲 Note that the output of Eq. 7 becomes the input to Eq. 8. Moreover,
this treatment makes no assumptions about the nature of the relation-
In practice, the FG is usually N2O. It is also possible to obtain the ship governing the interaction between the gas G and blood and is
following expression for V̇I: therefore always applicable under steady-state conditions. Further-
␭共FI ⫺ Fv៮兲 more, defining VA as the alveolar air volume, we can add the term
V̇I ⫽ V̇A ⫹ V̇A (4) d(FAVA)/dt to the right-hand side of Eq. 1, a term which represents the
共
␭共1 ⫺ FI兲 ⫹ V̇A ⁄ Q̇t 兲 change in the amount of gas in the lungs with time, to extend the
principle enunciated in Eqs. 7 and 8 to the non steady-state, indicating
Extension to a nonhomogenous lung with a continuous distribution that the principle is universally applicable to gas exchange in the lung.
of V̇/Q̇ ratios. With a log normal distribution of V̇/Q̇ ratios, Eq. 4 takes When the uptake of SG occurs in the presence of a soluble FG, we
the following form: therefore use the following sequence which mimics Eqs. 7 and 8: 0 ⫽
V̇I(x) ⫽ 再 V̇A
兹2␲
e⫺ 冉冊 ⁄2
x⫺
␴ 2
2 冎 both gases are introduced via the inspired alveolar ventilation to the
gas-exchanging unit; 1 ⫽ SG is then equilibrated with the blood flow
to the unit; and 2 ⫽ gas exchange is then completed by allowing the
⫹
再 V̇A
兹2␲ 关␭共1 ⫺ FI兲e冉x ⫺ 2 冊

␭共FI ⫺ Fv៮兲
␴ 2
⁄ 2 ⫹ 共V̇ A 兲
⁄ Q̇t e 冉冊 ⁄ 2兴
x⫹
␴ 2
2
冎 (5)
FG to equilibrate with the blood flow to the unit.
Splitting the gas exchange process into two steps in the gas-
exchanging unit makes it possible to evaluate the effect of the volume
contraction that occurs in that unit. This is similar to the technique
Here ␴ is the absolute value of the difference between the standard used by Rahn and Fenn to explain the divergence of the gas R lines on
deviation (SD) of the logarithm of ventilation per unit volume and the the O2-CO2 diagram (24). There the difference between V̇I and V̇A
SD of the logarithm of blood flow per unit volume; ␴ specifies the arises whenever the output of CO2 does not equal the uptake of O2 and
degree of mismatch between ventilation and blood flow (3). In awake the respiratory exchange ratio R, therefore, does not equal 1. In our
healthy adults, ␴ varies from 0.25 to 0.5; worsening of V̇/Q̇ matching case, the difference arises because of the uptake of the soluble first
is seen in disease states and is also known to occur soon after gas. In both cases the input to the single lung unit is quantified by
induction of anesthesia, with typical values of ␴ between 0.75 and specifying the composition of the inflowing blood and gas. The output
1.75 in healthy adults (4, 7, 13, 14, 25). is then uniquely determined once an equilibrium condition is identi-
Summed over the whole lung Eq. 5 provides a one-to-one relation- fied. In the case of the O2-CO2 system, a sufficient condition is the
ship between V̇I and FI and hence allows the calculation of the nomination of a particular value for R; in the case of our inert FG, it
inspired concentration of the FG associated with the desired total is the applicability of Henry’s law as soon as ␭ is specified. The
two-step model may next be extended, by analogy, to a multicom-
inspired alveolar ventilation and the accompanying value of ⌬V̇ ⫽ partment model and then in the limiting case to a lung model with a
共V̇I ⫺ V̇A兲 (11). continuous distribution of V̇/Q̇ ratios.
In Eq. 5, the term inside the first set of curly brackets generates the Because we are using N2O solely as a means of producing a
normal distribution of the expired ventilation while the term inside the volume change whose effects are to be studied, for convenience, we
second set of curly brackets generates the distribution of the first gas set Pv៮ to 0. Similarly, we set Pv៮ to 0 for the SG, since this maximizes
uptake. The distributions of V̇I, V̇A, and FG uptake may therefore be the SGE. Unless stated otherwise, FI is set at 0.7 for N2O, the highest
determined from Eq. 5. clinically effective concentration that can be used safely in the general
If a soluble second gas is present in low concentrations (we use a population without producing hypoxia (15). To mimic a high degree
concentration of 1% in the dry inspired gas mixture), it will also obey of V̇/Q̇ mismatch, we set ␴ ⫽ 2. To clarify the role of SG solubility,
J Appl Physiol • doi:10.1152/japplphysiol.00049.2020 • www.jap.org

ANALYSIS OF THE SECOND GAS EFFECT USING A TWO-STEP MODEL 1589
it is useful to select SGs with a wide range of solubilities in blood. We centrations of second gases as ␭ is increased. Alveolar ventilation V̇A
use sulfur hexafluoride (␭ ⫽ 0.0076), desflurane (␭ ⫽ 0.42), halo- is set at 4 L/min and pulmonary blood flow Q̇t is set at 5 L/min. All
thane (␭ ⫽ 2.4), diethyl-ether (␭ ⫽ 15.2), and acetone (␭ ⫽ 245) as programs were written in MATLAB. Further details of our technique
SGs in our study of gas exchange at the alveolar level. For ␭ ⬍ 10, gas are discussed elsewhere (10, 11). The equations used to generate the
exchange occurs largely in the alveoli. For 10 ⬍ ␭ ⬍ 100, Anderson figures in RESULTS are discussed further in APPENDIX B.
et al. have shown that there is a gradual transition from alveolar
exchange to airway exchange, whereas for ␭ ⬎ 100, gas exchange RESULTS
occurs almost exclusively in the airways (1, 2). Because exchange of
acetone falls in this last category, and it is in any case not delivered The results of our simulation are shown in Fig. 1, which
by inhalation for clinical purposes, our modeling of the SGE on depicts gas uptake as a two-step process for each gas under
acetone uptake in the alveolus is purely hypothetical and is included investigation. In each panel, gas uptake is plotted as a function
only to demonstrate the underlying trend in the SGE on trace con- of log10(V̇/Q̇) in the range [⫺3, 4]. The first step is shown in
Fig. 1. Uptake profile of sulfur hexafluoride

(A), desflurane (B), halothane (C), diethyl-
ether (D), and acetone (E). Step 1: distribu-
tion of each second gas (SG) comprising 1%
of the inspired gas mixture before (vertical
hatched region) and after (cyan-colored
area) equilibration with blood but before
N2O exchange is permitted. The distribution
of the absorbed gas is shown in yellow. Step
2: N2O uptake is now allowed to take place.
The red shading indicates the distribution of
the volume contraction that is to take place
in this step. This has been scaled to approx-
imately (1/70)th of its actual value so as not
to dwarf the volume of each SG. The distri-
bution of SG in the gas phase at the end of
step 1 acts as the starting point for the further
uptake in step 2 (vertical hatched region in
right-hand panel) that accompanies the con-
traction in volume producing the final distri-
bution of SG in the gas phase (cyan-colored
area above the x-axis). The extra uptake is
shown in yellow. The arrows in A and D
point to the sliver of SG uptake in step 2 for
sulfur hexafluoride and diethyl-ether. There
is virtually no extra uptake predicted in the
case of acetone.

the left-hand panel of each pair, the second step in the right-
hand panel. In the left-hand panel, exchange of 1% SG is
allowed to occur before the uptake of 70% N2O. The vertical
hatched region in each graph shows the distribution of the
inflowing SG. The portion of this gas transferred to blood is
shown in yellow, that remaining after uptake is shown in cyan.
The cyan-colored area in step 1 becomes the inflow of SG to
step 2.
N2O uptake is now permitted to take place in step 2 and is
accompanied by a contraction in volume proportional in mag-
nitude to the size of the red area. Because of the large
discrepancy between the concentrations of the N2O and SG
(70% vs. 1%), the red area is scaled to approximately (1/70)th
of the N2O uptake per minute. This area shows the distribution Fig. 3. Second gas (SG) uptake profile in step 2 of schematic mechanism
of the N2O uptake. shown in Fig. 1, as a function of log10(␭SG) with ␴ varied from 0 to 2 in
The contraction in volume produces an increase in the partial increments of 0.01. The N2O uptake is kept constant at 575 mL/min by
pressure of SG in the gas phase, leading to a further uptake of changing FI. The blue curve represents a value of ␴ ⫽ 2 (as in Fig. 2), and the
red curve represents a value of ␴ ⫽ 0. The direction from red to blue of a
SG by blood shown in yellow. The sequence is similar to that vertical line drawn at any given solubility indicates whether SG uptake in step
shown in step 1; however, the cyan-colored area in step 1 2 increases or decreases as ␴ is increased. Red and blue curves intersect at
becomes the starting point for SG uptake in step 2 and, points X and Y. To the right of Z, gas exchange occurs predominantly in the
therefore, forms the vertical hatched region in each right-hand airways. In this simulation, the points X, Y, and Z correspond to values of
␭ ⫽ 0.5, 10, and 100, respectively.
panel.
It immediately becomes obvious that, for each SG, the
yellow area in the right-hand panel depicts the SGE, which has
On the other hand, if the transfer of SG to blood in step 2 is
now been isolated from any other contributions to the alveolar
to be maximized, ␭SG should be as high as possible. These
uptake of SG. The largest transfer of SG occurs in Fig. 1B with
conflicting requirements for maximization suggest that, for any
desflurane. With sulfur hexafluoride and diethyl-ether, the
uptake in step 2 is virtually imperceptible (see arrows). In the given set of the other input variables (␭FG, V̇A, Q̇t, ␴, FI, and
case of acetone, there is negligible further uptake in step 2. Fv៮), there is a value of ␭SG for which the SGE is maximized. In
the sequence of gases shown in Fig. 1, this occurs with
DISCUSSION desflurane. In Fig. 2, we have drawn a graph of SG uptake
in step 2 vs. log10(␭SG), using the conditions that apply in
Having separated the uptake of SG into two components, the
Fig. 1. It can be seen that the maximum occurs for
first, which takes place at constant volume, the second associ-
log10(␭SG) ⫽ ⫺0.41, which corresponds to a value of 0.4 for
ated with the contraction of volume due to N2O uptake, it is
␭SG, consistent with the position of desflurane (␭ ⫽ 0.42) in
now possible to analyze the contributions of gas solubility and
the sequence shown in Fig. 1.
V̇/Q̇ mismatch to the SGE.
Role of ventilation-perfusion mismatch in relation to the
Role of the solubility of the second gas. For the SGE to be
second gas. It is now evident that V̇/Q̇ mismatch plays a
maximized, the amount of SG exposed to the contraction in net
similar role to that of ␭SG, needing to be as high as possible in
gas volume in step 2 must be as great as possible. This gas is
step 1 so as to conserve as much of the SG as possible in the
shown in cyan in each left-hand panel of Fig. 1. Maximization
first step but as low as possible in step 2 so as to facilitate
of the cyan-colored area will occur if the loss of SG to blood
maximum transfer of the SG to blood during the second step.
during the first step is minimized, i.e., if the solubility of SG in
We can therefore expect that, for each value of ␭SG, there is a
blood, ␭SG, is as low as possible.
value of ␴ that maximizes the transfer of SG in step 2.
Alignment vs. retention. In Fig. 3 we have graphed the SG
uptake in step 2 as a function of log10(␭SG) for values of ␴
ranging from 0 to 2. The contraction volume has been kept
constant and equal to that in Fig. 1 by varying the inspired
concentration of N2O. The red curve corresponds to the
situation when ␴ ⫽ 0; the blue curve corresponds to the
situation when ␴ ⫽ 2. The two curves cross at the points
labeled X and Y. To the left of X, an increase in V̇/Q̇
mismatch is associated with an increase in the amount of SG
transferred in step 2. Because there is already a significant
degree of SG retention in this solubility range (see Fig. 1A),
the likeliest cause of the increased SGE is the improved
alignment between the N2O uptake and the SG retained at
the end of step 1. To the right of point Y, an increase in V̇/Q̇
Fig. 2. Second gas (SG) uptake profile in step 2 of schematic mechanism mismatch is associated with a similar but smaller increase in
shown in Fig. 1, as a function of log10(␭SG) with FI ⫽ 0.7 for N2O and the amount of SG transferred in step 2. There can be little
␴ ⫽ 2. improvement in alignment here because of the wide sepa-

SGE is shown to accompany an increase in V̇/Q̇ mismatch,
but, since gas exchange occurs primarily in the airways for
␭ ⬎ 100 (1, 2), no V̇/Q̇ effect will be discernible clinically.
Location of gas uptake in the V̇/Q̇ spectrum. The location of
gas uptake and how it is affected by increases in V̇/Q̇ mismatch
influences the degree of overlap between the SG left after step
1 and the volume contraction in step 2. To understand where
gas uptake occurs in the V̇/Q̇ spectrum, we first examine the
effect of an increase in ␴ on the uptake-distribution of a gas
present in very low concentrations in the inspired gas mixture.
This is demonstrated for a relatively insoluble gas in Fig. 4A
using desflurane. When ␴ ⫽ 0, there is no mismatch, and the
uptake is distributed symmetrically about the y-axis (dark blue
curve). As the degree of V̇/Q̇ mismatch is increased, the uptake
profile shifts further and further into the low V̇/Q̇ zone. We can
describe desflurane uptake as perfusion-dependent when V̇/Q̇
mismatch increases. The situation for a highly soluble gas is
shown in Fig. 4B using halothane. It can be seen that the uptake
profile shifts further and further into the high V̇/Q̇ zone and is
thus ventilation-dependent.
It is possible for the uptake of a gas to be neither perfusion-
dependent nor ventilation-dependent as ␴ is increased. This is
shown in Fig. 4C and occurs when ␭ ⫽ V̇A/Q̇t which in our
case is equal to 0.8. N2O uptake is an example of perfusion-
dependence (␭ ⫽ 0.47). As FI increases, the uptake-distribu-
tion curve shifts to the left because the solubility of any inert
gas effectively becomes ␭共1 ⫺ FI兲. This is evident from Eq. 5
and has been reported previously (5, 27). Thus 70% N2O
behaves like a trace concentration of a gas with ␭ ⫽ 0.141. To
Fig. 4. Fractional uptake of 1% inspired concentration of gas as a function of
summarize, both the effective solubility, ␭⬘ ⫽ ␭共1 ⫺ FI兲, and
V̇/Q̇ mismatch for increasing values of ␴. A: desflurane (␭ ⫽ 0.42), B: the degree of V̇/Q̇ mismatch, ␴, affect uptake in each step
halothane (␭ ⫽ 2.4), and C: hypothetical gas with ␭ ⫽ V̇A/Q̇. ␴ ⫽ 0 (blue), according to Table 1.
␴ ⫽ 0.5 (red), ␴ ⫽ 1 (green), ␴ ⫽ 1.5 (purple), ␴ ⫽ 2 (magenta). The graphs Rising Pv៮N O with time will reduce the magnitude of the
in A and B become asymmetrical as ␴ is increased, so the vertical line in each 2
graph has been positioned at the median value of y. The uptake profile of effects modeled, but clinical studies have shown that the SGE
desflurane shifts to the left as ␴ increases, an example of perfusion-dependent persists at measurable levels for some time into maintenance
uptake; the uptake profile of halothane shifts to the right as ␴ increases, an phase anesthesia (8, 17, 21).
example of ventilation-dependent uptake (see direction of arrows in A and B).
Note that we have used the fraction of the total uptake as the variable on the Conclusion. We have investigated the alveolar transfer of
y-axis. The unit used on the x-axis is a function of log(V̇/Q̇) and is discussed gases in the lung using a two-step model, which allows the
further in APPENDIX A. contribution from net gas volume contraction, which occurs in
step 2, to be separated from other factors. The magnitude of the
ration of the retained gas and the volume contraction in second gas effect depends on sufficient inflow of the gas from
terms of their position on the x-axis (see Fig. 1E). This step 1 to step 2, which itself depends critically on how much
separation worsens as ␴ is increased (see Fig. 4 above and gas is retained after uptake in step 1.
associated discussion) so the increased SGE is probably due An additional factor is the proximity of the effective solu-
to an increase in the amount of SG retained at the end of step bilities of the first and second gases. This controls the extent to
1. Thus, we conclude that, as we move from left to right which the SG retained at the end of step 1 is exposed to the
along the solubility axis, the importance of correct align- contraction in gas volume that takes place in step 2. The
ment wanes while that of retention increases. In the region information in Table 1 allows us to conclude that both solu-
between X and Y, these processes are not powerful enough bility and V̇/Q̇ mismatch act to increase uptake in step 2 by
to sustain an increase in the SGE with an increase in V̇/Q̇ decreasing uptake in step 1. This provides an explanation of
mismatch. To the right of point Z in Fig. 3, an increase in the how V̇/Q̇ mismatch and low gas solubility combine to increase
Table 1. Factors affecting gas uptake in each step of a mechanistic model of gas exchange
Factor Under Consideration Step 1 Step 2
Low effective solubility Favors retention in gas phase Impedes gas transfer to blood
High effective solubility Favors loss of gas to blood Favors gas transfer to blood
Low-degree V̇/Q̇ mismatch Favors loss of gas to blood Favors gas transfer to blood
High-degree V̇/Q̇ mismatch Favors retention in gas phase Impedes gas transfer to blood

gas uptake during N2O anesthesia in a manner that is entirely APPENDIX B

compatible with existing concepts of gas exchange.
Derivation of distributions of inflow, uptake, retention, and outflow
for use in diagrams. We start by defining the interval on the x-axis to
APPENDIX A be partitioned into n equal intervals. Because we have specified that Pv៮
equals 0 for each gas under consideration, the only input we need to
The Colburn equations. The probability density functions for ven- consider is gas being brought into the lung by the inspired ventilation.
tilation (V̇) and blood flow (Q̇) derived by Colburn et al. (3) are as Therefore, an acceptable interval is where 共␮ ⫺ 3兲 ⱕ x ⱕ 共␮ ⫹ 3兲. If
follows: contributions from mixed venous blood need to be considered, the
left-hand limit must be extended to 共⫺␮ ⫺ 3兲. If we restrict ␴ to the
V̇A 2⁄2 interval [0, 2], a partition on the interval [⫺4, 4] can be used throughout.
V̇ ⫽ e⫺x (A1) Choosing n ⫽ 800 will then provide 101 points in each unit subinterval,
兹2␲ e.g., from ⫺4 to ⫺3, each of the n intervals will be 0.01 units wide. This
will generally provide sufficient coverage. We now have a partition {xk}
Q̇t
Q̇ ⫽ e⫺(x ⫺ ␴)
2⁄2
(A2) of the interval [⫺4, 4] with n ⫹ 1 points such that ⫺ 4 ⬍ x1 ⬍
兹2␲ x2. . .xn ⬍ xn⫹1 ⬍ 4 and define the vector x ⫽ 关x1 . . . xn⫹1兴. We next
distinguish the different stages of gas uptake as follows: 0 ⫽ initial
Here V̇A and Q̇t are the total expired alveolar ventilation and total introduction of the gas G under consideration; 1 ⫽ situation at the end of
pulmonary blood flow, respectively, and ␴ is as previously defined in step 1; and 2 ⫽ situation at the end of step 2.
relation to Eq. 5. Dividing Eq. A1 by Eq. A2 and solving for x we The change in any parameter in step 1 is then indicated by the
obtain: abbreviation (01) after the symbol for that parameter, and that in step
2 by the abbreviation (12). The total change in steps 1 and 2 is
␴ 1
x⫽
2
⫺
␴
关log共V̇ ⁄ Q̇兲 ⫺ log共V̇A ⁄ Q̇t兲兴 (A3)
indicated by the abbreviation (02).
For each xk in x we define the following parameters and the
associated vectors:
The inspired alveolar gas flow:
Defining v ⫽ V̇ ⁄ V̇A as the fractional ventilation and q ⫽ Q̇ ⁄ Q̇t as the
fractional blood flow, Eq. A3 simplifies to:
␴ log共v ⁄ q兲
V̇Ik ⫽ 再 V̇A
兹2␲
e⫺(xk ⫺ ␮)
2⁄2
冎
再冎
x⫽ ⫺ (A4)
2 ␴ V̇A ␭(FI ⫺ Fv៮)
⫹
兹2␲ 关␭(1 ⫺ FI)e共x 兴
2⁄2
k ⫺ ␮兲 ⁄ 2 ⫹ (V̇A ⁄ Q̇t)e(xk ⫹ ␮)
2
Standard graphs of ventilation and blood flow as functions of log(V̇/ (B1)
Q̇) change size and shape as ␴ is increased; the graphs for ventilation
and blood flow move away from each other in proportion to the square 关
V̇I ⫽ V̇I1 . . . V̇In⫹1 兴
of ␴, the spread or dispersion of the graphs increases in proportion to The expired alveolar gas flow:
␴, and the height of each graph decreases in proportion to ␴. An
alternative, suggested by Eq. A4, is to use (1/␴)log(v/q) as the V̇A 2⁄2
independent variable (10). V̇Ak ⫽ e⫺(xk ⫺ ␮)
Using this substitution in Eqs. A1 and A2, and letting ␮ ⫽ ␴ ⁄ 2, 兹2␲ (B2)
we obtain the following alternative expressions for V̇ and Q̇:
关
V̇A ⫽ V̇A1. . .V̇An⫹1 兴
V̇A 2⁄2 The pulmonary capillary blood flow:
V̇ ⫽ e⫺(x ⫺ ␮) (A5)
兹2␲ Q̇t 2⁄2
Q̇k ⫽ e⫺(xk ⫹ ␮)
Q̇ ⫽
Q̇t
e⫺(x ⫹ ␮)
2⁄2
(A6)
兹2␲ (B3)
兹2␲ 关
Q̇ ⫽ Q̇1 . . . Q̇n⫹1 兴
Drawn in this way, the graphs for ventilation and blood flow are Uptake of gas G in step 1:
always exactly the same size and shape with amplitude 1 ⁄ 兹2␲ and
SD ⫽ 1. When ␴ ⫽ 0, there is no V̇/Q̇ mismatch, and the two curves ␭V̇Ik(FI ⫺ Fv៮)
V̇G(01)k ⫽
overlap completely. As ␴ increases, the curves move apart, with the ␭ ⫹ (V̇Ik ⁄ Q̇tk) (B4)
means for ventilation and blood flow located exactly ␴/2 units to
the right and left of the y-axis (x ⫽ 0 line), respectively. Using 关
V̇G(01) ⫽ V̇G(01)1 . . . V̇G(01)n⫹1 兴
(1/␴)log(v/q) as the variable on the x-axis therefore provides a stable
platform for studying the uptake of the first and second gases when ␴ Combined uptake of gas G in steps 1 and 2:
is allowed to vary. This alternative unit has been used on the x-axis in
V̇A ␭(FI ⫺ Fv៮)
Fig. 4. However, because ␴ is kept constant and equal to 2 in Fig. 1, V̇G(02) ⫽
兹2␲ 兵␭(1 ⫺ FI)e(x 其
2 2⁄2
it is just as convenient to plot gas uptake in this figure as a function k ⫺ ␮) ⁄ 2 ⫹ (V̇A ⁄ Q̇t)e(xk ⫹ ␮)
共兲
of log10 V̇ ⁄ Q̇ .
关
V̇G(02) ⫽ V̇G(02)1 . . . V̇G(02)n⫹1 兴
Using the chain rule, it is possible to show that changes in the
distributions as a function of log10(V̇/Q̇) are mirrored by changes of (B5)
similar sign when the distributions are expressed as a function of
(1/␴)log(v/q). Hence, the direction of the arrows in Fig. 4 are the same The inflow of gas G in step 1 is obtained by multiplying the vector V̇I
as those that would apply if we were plotting against log10(V̇/Q̇). by the fractional concentration of G in inspired gas i.e., V̇IG共01兲 ⫽

10. Korman B, Dash RK, Peyton PJ. Effect of net gas volume changes on
V̇IFIG, The gas retained at the end of step 1 becomes the outflow of the
alveolar and arterial gas partial pressures in the presence of ventilation-
gas G at the end of step 1: V̇AG共01兲⫽V̇IG共01兲 ⫺ V̇G共01兲. This in turn perfusion mismatch. J Appl Physiol (1985) 126: 558 –568, 2019. doi:10.
becomes the inflow of gas G in step 2. The uptake of gas G in step 2 1152/japplphysiol.00689.2018.
11. Korman B, Mapleson WW. Concentration and second gas effects: can
is given by V̇G共12兲⫽V̇G共02兲 ⫺ V̇G共01兲. Thus, by using vector algebra, the accepted explanation be improved? Br J Anaesth 78: 618 –625, 1997.
it is possible to obtain the distribution of any required parameter. The doi:10.1093/bja/78.5.618.
total uptake of gas over the whole range of x values may be obtained 12. Lundh R, Hedenstierna G. Ventilation-perfusion relationships during
by integrating the relevant expression using the integral function of halothane anaesthesia and mechanical ventilation. Effects of varying
MATLAB, but this involves such a complicated expression for the inspired oxygen concentration. Acta Anaesthesiol Scand 28: 191–198,
integrand in the case of Eq. B4 that the much simpler option of 1984. doi:10.1111/j.1399-6576.1984.tb02039.x.
applying the trapezoidal rule (9) has been used instead here and for 13. Lundh R, Hedenstierna G. Ventilation-perfusion relationships during
other calculations, where appropriate. anaesthesia and abdominal surgery. Acta Anaesthesiol Scand 27: 167–173,
1983. doi:10.1111/j.1399-6576.1983.tb01929.x.
DISCLOSURES 15. Magnusson L, Spahn DR. New concepts of atelectasis during general
anaesthesia. Br J Anaesth 91: 61–72, 2003. doi:10.1093/bja/aeg085.
No conflicts of interest, financial or otherwise, are declared by the authors.
16. Neufeld GR, Williams JJ, Klineberg PL, Marshall BE. Inert gas a-A
AUTHOR CONTRIBUTIONS differences: a direct reflection of V/Q distribution. J Appl Physiol 44:
277–283, 1978. doi:10.1152/jappl.1978.44.2.277.
B.K. conceived and designed research; B.K. performed experiments; B.K. 17. Peyton PJ, Horriat M, Robinson GJ, Pierce R, Thompson BR. Mag-
analyzed data; B.K. and P.J.P. interpreted results of experiments; B.K. and nitude of the second gas effect on arterial sevoflurane partial pressure.
R.K.D. prepared figures; B.K. drafted manuscript; B.K., R.K.D., and P.J.P. Anesthesiology 108: 381–387, 2008. doi:10.1097/ALN.0b013e318164caf3.
edited and revised manuscript; B.K., R.K.D., and P.J.P. approved final version 18. Peyton PJ, Robinson GJ, Thompson B. Effect of ventilation-perfusion
of manuscript. inhomogeneity and N2O on oxygenation: physiological modeling of gas
exchange. J Appl Physiol (1985) 91: 17–25, 2001. doi:10.1152/jappl.2001.
REFERENCES
91.1.17.
1. Anderson JC, Babb AL, Hlastala MP. Modeling soluble gas exchange 19. Peyton PJ, Robinson GJ, Thompson B. Ventilation-perfusion inhomo-
in the airways and alveoli. Ann Biomed Eng 31: 1402–1422, 2003. geneity increases gas uptake in anesthesia: computer modeling of gas
doi:10.1114/1.1630600. exchange. J Appl Physiol (1985) 91: 10 –16, 2001. doi:10.1152/jappl.
2. Anderson JC, Hlastala MP. Breath tests and airway gas exchange. Pulm 2001.91.1.10.
Pharmacol Ther 20: 112–117, 2007. doi:10.1016/j.pupt.2005.12.002. 20. Peyton PJ, Robinson GJ, Thompson B. Ventilation-perfusion inhomo-
3. Colburn WE Jr, Evans JW, West JB. Analysis of effect of the solubility geneity increases gas uptake: theoretical modeling of gas exchange. J Appl
on gas exchange in nonhomogenous lungs. J Appl Physiol 37: 547–551, Physiol (1985) 91: 3–9, 2001. doi:10.1152/jappl.2001.91.1.3.
1974. doi:10.1152/jappl.1974.37.4.547. 21. Peyton PJ, Stuart-Andrews C, Deo K, Strahan F, Robinson GJ,
4. Dueck R, Young I, Clausen J, Wagner PD. Altered distribution of Thompson BR, Pierce R. Persisting concentrating and second gas effects
pulmonary ventilation and blood flow following induction of inhalation on oxygenation during N2O anaesthesia. Anaesthesia 61: 322–329, 2006.
anesthesia. Anesthesiology 52: 113–125, 1980. doi:10.1097/00000542- doi:10.1111/j.1365-2044.2006.04579.x.
198002000-00004. 23. Peyton PJ, Fortuin M, Robinson GJB, Stuart-Andrews C, Pierce R,
5. Eger EI II, Smith RA, Koblin DD. The concentration effect can be Thompson BR. The rate of alveolar-capillary uptake of sevoflurane and
mimicked by a decrease in blood solubility. Anesthesiology 49: 282–284, nitrous oxide following anaesthetic induction. Anaesthesia 63: 358 –363,
1978. doi:10.1097/00000542-197810000-00012. 2008. doi:10.1111/j.1365-2044.2007.05355.x.
6. Epstein RM, Rackow H, Salanitre E, Wolf GL. Influence of the 24. Rahn H, Fenn WO. A Graphical Analysis of the Respiratory Gas
concentration effect on the uptake of anesthetic mixtures the second
Exchange. Washington, D.C.: The American Physiological Society, 1955,
gas effect. Anesthesiology 25: 364 –371, 1964. doi:10.1097/00000542-
p. 39.
196405000-00015.
14. Hedenstierna G, Lundh R, Johansson H. Alveolar stability during 25. Rehder K, Knopp TJ, Sessler AD, Didier EP. Ventilation-perfusion
anaesthesia for reconstructive vascular surgery in the leg. Acta Anaesthe- relationship in young healthy awake and anesthetized-paralyzed man. J
siol Scand 27: 26 –34, 1983. doi:10.1111/j.1399-6576.1983.tb01900.x. Appl Physiol 47: 745–753, 1979. doi:10.1152/jappl.1979.47.4.745.
7. Hendrickx JF, Carette R, Lemmens HJM, De Wolf AM. Large volume 26. Stoelting RK, Eger EI II. An additional explanation for the second gas
N2O uptake alone does not explain the second gas effect of N2O on effect: a concentrating effect. Anesthesiology 30: 273–277, 1969. doi:10.
sevoflurane during constant inspired ventilation. Br J Anaesth 96: 391– 1097/00000542-196903000-00007.
395, 2006. doi:10.1093/bja/ael008. 27. West JB, Wagner PD, Derks CM. Gas exchange in distributions of
8. Karris S. Numerical Analysis Using MATLAB and Excel. New York, NY: VA-Q ratios: partial pressure-solubility diagram. J Appl Physiol 37:
Orchard Publications, 2007, p. 391–393. 533–540, 1974. doi:10.1152/jappl.1974.37.4.533.
9. Korman B, Dash RK, Peyton PJ. Can mathematical modeling explain 28. West JB, Wagner PD. Pulmonary gas exchange. In: Bioengineering
the measured magnitude of the second gas effect? Anesthesiology 128: Aspects of the Lung, edited by West JB. New York: Marcel Dekker, 1977,
1075–1083, 2018. doi:10.1097/ALN.0000000000002131. p. 378 –382.

THESIS DOCTOR OF PHILOSOPHY KORMAN Bengt 2020

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

THESIS DOCTOR OF PHILOSOPHY KORMAN Bengt 2020

Uploaded by

Copyright:

Available Formats

MODELLING OF GAS EXCHANGE

IN THE LUNG DURING NITROUS

Dr Bengt Korman OAM

Submitted in fulfilment of the requirements for the degree of

Medical School, Faculty of Health and Medical Sciences

Dept. of Mathematics and Statistics, Faculty of Engineering and Mathematical Sciences

The University of Western Australia 2020

Dr Barry Marshall AC, 2005 Nobel Laureate in Medicine & Physiology

Dr Stephan Schug, Professor of Anaesthesia and Pain Medicine, UWA

Dr Snezhana Abarzhi, Professor of Applied Mathematics, UWA

Dr Philip Peyton, Professor of Anaesthesia & Pain Medicine, University of Melbourne

Last revised 21 September 2020

Anaesthetic uptake, concentration effect, mathematical modelling, second gas

effect, ventilation-perfusion mismatch

termed a “homogeneous lung”. Thus, a fundamental principle of respiratory

In 2001, using computer modelling, it was predicted that increased

ventilation-perfusion mismatch could improve gas uptake during nitrous oxide

anaesthesia. The prediction seemed to be confirmed by subsequent experimental

enunciated above. Therefore, the hypothesis for this research was:

Increased ventilation-perfusion mismatch per se cannot improve gas exchange.

formal mathematical analysis developed in the Departments of Medicine and

Mathematics at The University of California, San Diego in the 1970s. Their

were obtained using MATLAB, a programming language with powerful problem-

To independently replicate the theoretical basis for this strange new

phenomenon. This was necessary to ensure that the previous computer

To investigate further the effect of nitrous oxide anaesthesia on other gases

present using partial pressure-solubility diagrams. This was necessary because

To find a plausible explanation of how ventilation-perfusion mismatch can act

to improve gas transfer without violating the fundamental principle of gas

exchange in the lung.

confirmed in part one.

introduction of electronic depth-of-anaesthesia monitors. I also showed that many

To explain these findings, I appealed to a previously unrecognized property

correction. The strange findings referred to above, always accompanied the

were shown to be a product of step two. Accordingly, their magnitude depended on

retention at the end of step one, an increase in ventilation-perfusion mismatch exerts

The mysterious phenomena first described in 2001 challenged a fundamental

Figure 2.1. The concentration effect. ......................................................................... 13

List of Figures xiii

xiv List of Figures

Table 5.1. The nitrous oxide Washin Ratio…………………………………..66

V Gas volume in general

Symbols for gas phase:

Symbols for blood:

Other context-specific abbreviations are defined as the need arises.

List of Abbreviations xvii

The models presented in this thesis contain many assumptions. Most of

of their complicated interactions with haemoglobin

2. All other gases obey Henry’s Law

3. Unless stated otherwise, steady-state conditions apply

4. The functional residual capacity is assumed to remain constant

during gas exchange

5. Absorption atelectasis and hypoxic vasoconstriction are not

These assumptions are discussed in further detail in relevant sections of the

xviii Assumptions made in Modelling

Details of the work:

Date: Date: 10/14/2019

Details of the work:

Date: Date: 10/14/2019

Details of the work:

Date: 17/03/2020 Date: 03/17/2020

written by Professor Ross Kennedy may be viewed at the ANESTHESIOLOGY

128: 1053-1054, 2018 (https://anesthesiology.pubs.asahq.org/article.aspx?articleid=2677923).