Eur Respir J 1990, 3, 469--482 REVIEW
Clinical relevance of ventilation-perfusion inequality determined by inert gas elimination
R. Rodriguez-Roisin, P.O. Wagner
Clinical relevance of venlilation-perfusion inequality determined by inert gas Depts of Medicine, Scrvei de Pncumologia, elimination. R. Rodriguez-Roisin, PD. Wagner. Hospital Clinic, Facultat de Medicina, Universitat ABSTRACT: The tirst part of this review deals with the basic mechanisms de Barcelona, Barcelona, Spain, and Section of and factors determining hypoxaemia and hypercapnia and the different Physiology, School of Medicine, University of La approaches used In cllnlc:tl practice and In clinical research to assess the California at San Diego, Jolla, California, USA. presence of venlllatlon-perfuslon mlsmatchlng1 shunt and diffusion llmlta Correspondence: R. Rodriguez-Roisin, Servei de tlon for OA)'ge n, and more spcclncally the multiple Inert gns elim ination Pncumologia, Hospital Clinic, Villarroel 170, 08036- technique (l\1 IGET), In pulmonary medicine. 1'he second part reviews three Barcelona, Spain. different respiratory disorders where the complex Interplay between Intrapul Keywords: Bronchial asthma; gas exchange; monary and extrapulmonury factors regulating oxygen are essentially Inter haemodialysis; hypoxaemia; multiple inert gas preted th r ough the results afforded by the MlGET over the last decade. The elimination technique; pulmonary embolism. gas exchange response to bronchodilators In bronchial asthma, an airway Received: March 8, 1989; accepted after revision disease, and then the mnjor determinants governing abnormal gas exchange August 10, 1989. In acute pulmonary embolism1 a pulmonary vascular disorder, and during haemodialysis, a respiratory entity of extrapulmonary origin, are succes Supported by Grants from the Joint US-Spain sively explored In the light of the Inert gas method. Committee (CCA 8309185), the National Heart, Lung and Blood Institutes (HL 17731-11) (USA), Eur Respir 1., 1990, 3, 469-482 and la Direccion General de Investigaci6n Cientlfica y Tecnica (DG ICYT) (PA85-0016) (Spain).
Hypoxaemia and hypercapnia in lead to an increase in arterial carbon dioxide tension
respiratory disease (Paco2).With hypoven~ilatfon, hypercapnia will always be present, while with V A/Q inequality, hypercapnia may Most respiratory diseases, both acute and chronic, have or may not be present [1]. depending upon the underly multiple physical effects upon the lungs. Mechanical ing individual ventilatory conditions. As shown in table properties are particularly affected by the obstructive and 1, Lhe most common, and quantitatively the most signifi restrictive disease states such as chronic obstructive lung cant, fac~or ~ontributing to abnonnal gas exchange is disease (COPD), bronchial asthma, and interstitial altered VA/Q relationships. Table 2 illustrates those pulmonary fibrosis. These, and other diseases, also affect intrapulmonary factors that contribute to hypoxaemia as other components of the lung such as the circulation and well as those extrapulmonary factors that, either directly the blood:gas barrier. Thus, pulmonary vascular obstruc or indirectly (Lhrough their effect on mixed venous oxygen tion and destruction and/or obliteration of pulmonary tension (Pvoz)), can also affect arterial oxygen tension ). blood vessels in COPD, asthma, and pulmonary fibrosis (Pao2 Total alveolar ventilation is considered here an affect both bloodOow and its distribution within the lung. extrapulmonary factor because it is primarily the result As a result, most lung diseases alter the distribution of of tidal volume (less series dead space-dead space ventilation, of bloodOow, or of both, and lead to abnor common to more than one YA/Q unit) and frequency, malities in pulmonary gas exchange. which are set by extrapulmonary breathing control Alveolar hypoventilation, shunt, diffusion limitation for mechanisms. Not shown in table 2 is the basis for how oxygen and ventilation-perfusion (VA/Q) inequality are extrapulmonary factors (other than ventilation and ) the four classical physiological mechanisms by which inspired Po2 may contribute to hypoxaemia. This is the such structural and functional changes due to lung level of mixed venous Po2 (Pvoz), a well-known concept, disease alter gas exchange. Table 1 lists the most com yet often overlooked in clinical settings. A decreased mon kinds of acute and chronic respiratory disorders that Pvo2 is the final common pathway by which the extrapul affect pulmonary gas exchange and show which of these monary factors other than inspired Po~ and ventilation mechanisms are most important in each disease state. lead to arterial hypoxaemia. Thus, mtxed venous hy This table also indicates typical arterial blood gas values poxaemia may result from a low cardiac output, an in these diseases. While all four causes of hypoxaemia increased 0 2 uptake or a decreased blood 0 2 content due can be important on different occasions, it is uncommon to several factors (altered haemoglobin concentration, for shunt or 0 2 diffusion limitation in the lung to temperature, Pw pH). In the presence of intrapulmonary 470 R. RODRIGUEZ-ROIS!N, P.D. WAGNER
Table 1. - Mechanisms of altered gas exchange in common pulmonary diseases
Typical Typical VA/Q Shunt Hypo- orl d~ffu.sion ventilation Disease Pao2 Paco2 inequality umtat10n COPD Type A 70 35 Yes No No No COPD Type B 50 50 Yes No Yes No Asthma 65 35 Yes No No No Interstitial Fibrosis 60 35 Yes Yes No Yes Pneumonia 50 30 Yes Yes No No ARDS 40 30 Yes Yes No No Pulmonary Embolism 70 30 Yes No• No No
COPD: Chronic obstructive pulmonary disease, types A and B according to Burrows' classification; ARDS: adult
resp~ratory distress syndrome; •: only at a late stage; Pao2 : arterial oxygen tension; Paco2: arterial carbon dioxide tensiOn; VA/Q: ventilation-perfusion J istribution; values are given in mmHg, breathing room air. abnormalities of gas exchange, Pvo2 can have a surpris Traditional ?1PP.roaches for measuring • V AIQ inequality ingly large effect on the resultant Pao2 For example, Pao2 in bronchial asthma often exceeds that of diffuse P.Ul~onary fibrosis, yet at the same time the amount of Topographical radioactive tracer-based approaches, VA/Q mismatching may be more severe in asthma. T'his namely ventilation-perfusion scans, have long been (and , are still) available to the clinician and serve extremely is explained by the higher cardiac output, hence Pvo2 usually seen in asthma [2, 3). useful functions in specific clinical situations such as The most important clinical int.r:apqlmonary factors suspected pulmonary embolism. However, when used in patients with generalized lung disease, such as COPD or determining Pao2 are shunL and V11/Q mismatching. Among the extrapulmonary ones, inspired 0 fraction, asthma, the results often grossly underestimate the degree 1 of ventilation/bloodflow mismatching actually present. total alveolar ventilation, cardiac output and 0 2 consump tion are viewed as the most innuential. Alternatively, This is easily demonstrated by the failure to accouql f9r Paco may also be determined by extrapulmonary the observed hypoxaemia on the basis of the V A/Q 2 inequality measured by s.uch external radioactive factors, due to changes in C02 production, alveolar ventilation, or in acid-base status. methods. The degree of VA/Q abnormality measured by topographic techniques is, thus, much less than that Table 2.- Intrapulmonary and extrapulmonary factors that actually present within the lung. The limited spatial contribute to hypoxaemia resolution of external counting devices leads to underes timation of the intra-regional VA/Q mismatching by Intrapulmonary Ex trapulmonary the isotopic methods, although this is improving nowa VA/Q Inequality Reduced: Total ventilation days with technical advances in equipment. The signal is not strictly proportional to bloodflow or alveolar vcntila· Shunt Cardiac output lion, but also reflects the distribution of the lung volume. Moreover, this can be the cause of an underestimation 0 Diffusion Inspired Po 1 2 of the inter-regional VA/Q mismatch. In the absence of direct measurements of cardiac output and ventilation, Limitation Haemoglobin isotopic approaches only give relative 11/Q distribution concentration V because the global VA/Q relationship is assumed to be equal to one. This approximation also leads to underes P 50 of 0 2 dissociation curve timation of the global VA/Q inequality. Unfortunately, use of .the respiratory gases 0 and CO Increased: 0 consumption 2 as tools for assessing VA/Q mismatching, leads some: pH times to unsatisfactory results that can be difficult to
VA/t).: vcntilation-pcrfusion ; Po : oxygen tension. interpret. There are several reasons for this. The classical 2 indices that are based upon the physiologic gases, such It should be apparent from table 1 that the most as vcn?us admixture 0>re,athing room air) or shunt (upon prominent of the pathophysiology of gas exchange in b~ea th tng oxygen) (Qs/QT), physiological dead space pulmonary disease is abnormal V A/Q relationships. A (Vo/VT) and alveolar-arterial 0 2 partial pressure differ rational approach to therapy and to understanding of the ences (P(A-a)oz), are all "as if' indicators of the degree underly~ng pathogenetic. m~chanisms, therefore, requires of heterogeneity. Venous admixture represents poorly the abthty to quantify V11/Q relationships in the lung. oxygenated blood which passes through the lung. It is VENTILATION-PERFUSION INEQUALITY 471
Lhat shunt which alone would explain all of the observed potential for extending FARm's approach (9] by using a hypoxaemia. Similarly, physiological dead space, another larger number of inert gases, and developed the so-called common parameter utilized to describe inefficiency of multiple inert gas elimination technique (MlGET), as a gas exchange, is that dead space which alone would means of obtaif!ing even more information about the explain the differences between arterial and mixed distribution of VA/Q relationships. It became evident expired Pco . It is composed of anatomical (airway) dead that fairly complex distributions of YA/Q ratios could be space and afvcolar (effective) dead space resulting from measured with as few as six inert gases [10]. Using mass V A/()_ heterogeneity. Neither venous .adrpixture nor balance equations, the fractional elimination of inert gases physiological dead space can separate VA/Q .ine!luality of widely differing solubility undergoing steady state from Lruc iny-ap~lmonary shunt in the case of Qs/Qr, nor elimination from several th~or~tical lung models changes in VA/Q abnormalit!es ~rom changes in airway containing different degrees of VA/Q mismatch was cal dead space in the case of Vo/Y.r. They are "lumped culated. Algorithms were devised which would take the parameter" estimates based upon simple two calculated elimination da~. a~d by using a smoothing compartment models of the lung [4]. Thus they are constraint [111 produce a VA/Q ratio distribution com difficult to interpret in terms of actual respiratory patible with the particular set of data in t11e presence of pathophysiology. In addition, these terms are highly experimental error. It was found that with ~ix gases it dependent upon the inflowing mixed venous blood:gas was possible to relatively accurately recover VA/Q distri composition. Generally, assumptions must be made about butions comaining up to three separate populations of mixed venous blood, and unless they can be independ lung units, even in the presence of realistic levels of ently verified by direct pulmonary arterial catheteriza noise. ti on, not always done in the clinical setting, the It soon became apparent that not only could the arte magnitude of the indices calculated can be considerably rial partial pressures of the eliminated gases be used in in error, particularly when following one patient over this manner to reconstruct the pulmonary perfusion time. Finally, each of these indices is sensitive to global distribution, but in a very symmetrical way, the mixed variables such as total alveolar ventilation, total cardiac expired partial pressures could be used to calculate the distribution of ventilation. Ultimately, arterial and mixed output, and inspired 0 1 fraction. When any of Lhese three parameters is altered, and this often happens in the criti expired data were combined in order to produce an inter cal care setting especially, changes in the indices will nally consistent combined distribution of ventilation and occur even if the actual amount of intrapulmonary bloodflow [11]. Figure 1 illusLrates the principles of the abnormality remains completely unaltered [4]. This adds technique showing, for four different types of VA/Q ratio another layer of confusion upon any interpretative scheme distribution, the corresponding relationships between inert ~f LJ:lese indices [5). Thus quantitation of the degree of gas elimination and solubility during steady state gas V A/Q mismatching has been a constant goal ever since exchange. From this figure, striking changes in the inert the initial three-compartment analysis developed almost gas eliminatipn ~ata are observed with significant altera simultaneously by Ft:NN et al. [6) and RILEY and tions in the VA/Q distribution pattern. Figure 1 is drawn CouRNAND [7] in the mid-1940's. Consequently, some based only on the steady state equations for gas exchange alternative means for assessing ventilation/bloodflow ine as defined by I a<.ttCOnt 1.0 normal .8 (\ •6 • .4 • .2 SF6 0 J \ .001 .1 10 1000 0 .001 10 1000 1.0 hyperventilation .8 " •6 • c .4 .2 • c .2 ) -Cll 0 \ 3: 0 -...Cll .001 .1 10 1000 0 0.001 .1 10 1000 ;: 'tJ Ill 0 cu 0 Cl m t:: 1.0 low VA!O areas cCll .8 .6 • . 4 • .1 10 1000 0 0.001 .1 10 1000 1.0 shunt .8 .6 • .4 • .2 0~~~· ..~ ·· -· ~~--~--~--~~ .001 .1 10 1000 0 0.001 .1 10 1000 Partition coefficient Ventilation/perfusion ratio Fig.l. - Retention (arteriaVmixed ven90s concentration ratio) of 6 inert gases plo!ted against !heir blood:gas partition coefficient (left panels) for four different vcntilation-perfuson (V A/Q) distributions shown in the righthand panels. The dotted lines on !he retention plot indicate the normal retention curve of !he uppermost panel, and !he arrows highlight !he nature of !he change in shape or position of !he retention curve <:::~used by !he particular distribution shown on !he right. No!ice !hat large alterations in retention data are produced by these distributions and, in particular, how distributions with low VA/Q areas and shunt lead to quite different retention-partition coefficient curves. measured during quiet breathing in mixed expired gas, Procedure arterial blood, and pulmonary arterial blood. Minute ventilation is measured together with the partition coef The experimental requirements for the method are ficients of the six gases in the patient's blood. If pulmo relatively simple. A mixture of the six gases is dissolved nary arterial sam!Jles cannot be obtained, mixed venous either in saline or dextrose which is then infused inert gas levels can be calculated by mass balance from intravenously into a peripheral vein at a rate of about arterial and mixed expired levels and cardiac output, 3 ml·min·1 (for a resting human being). After allowing minute ventilation and partition coefficients, as described some 30 min or so for the inert gases to achieve a steady below. These data can be used in addition to compute state within the lungs, the inert gas concentrations are cardiac output by mass balance for each gas in the mixture VENTILATION-PERFUSION INEQUALITY 473 and a mean cardiac output can be obtained. Inert gas A second alternative allows one to avoid even arterial concentrations are measured by gas chromatography after blood sampling. In this approach, the inert gases are extraction of gases from blood in gas-tight syringes [13]. infused in solution into the peripheral vein of one arm, The results can be expressed in a number of different and blood is sampled from a peripheral vein of the -.yay$. A pictorial representation of a multi-compartment opposite arm at the same time as making expired gas V A/Q distribution that fits the data can be drawn (fig. 2). measurements [19]. This technique, by avoiding both Total values of bloodflow and ventilation in defmed pulmonary arterial and systemic arterial catheterization, regions of the VA/Q spectrum can be given, moments is ideally suited for repetitive measurements over of the recovered distribution can be calculated (only the extended periods of time (for instance, in patients with first and second moments are routinely used), and bronchial asthma where questions arise about variability arterial-alveolar differences for the inert gases can also of ventilation/bloodflow mismatching during the course be c;:alcplated and used as indirect estimates of the degree of treatment). This approach, however, is restricted to of VA/Q mismatch. The second moment of the distribu situations of long-tenn (at least 90 min) steady state tion on a logarithmic scale, i.e. its dispersion, can be conditions and cannot follow rapid changes in function determined experimentally with a coefficient of variation that might occur over the time course of a few minutes. of between 8-9% for a single measurement. In the usual The rationale behind this approach is that the inert gases scheme, duplicate measurements are made, and this are not metabolized in the tissues of the hand, therefore, reduces that coefficient of variation by the square root of after a period of equilibration, the peripheral venous blood two to approximately 6% [16]. will have the same concentrations of each inert gas as will the inflowing arterial blood. This approach has 30'/, recently been validated by comparing it with the more t invasive one of using arterial blood in different series of i:: Dead space ....E t5 asthmatics followed weekly [20] . Resolution and advantages of the M/GET The technique allows one, for the first time, to obtain a close approximation to the functional distribution of VA/t) ratios within the lung. It is, however, a techni cally demanding approach and requires considerable theo retical knowledge for its proper interpretation and utilization. Its particular strengths include the tracer nature 100 100.0 of the approach, such that the mythod itself does not alter VAIQ ratio the physiology of gas exchange in the lung (the inert gas Fig. 2. - Example of the distribution of ventilation (o) and bloodflow levels are of the order of a few ppm). This should be (•) against ventilation-perfusion ('Q;JQ) ratio on a log scale in a resting contrasted with techniques in which inspired oxygen young individual, breathing room air. Note that both distributions are symmetrical. are appropriately positioned about a VA/0. ratio of 1.0 concentration is manipulated in order to gain insight into (1st moment); and there is little dispersion (narrowness) on each curve VA/Q relationships. Here there is every reason to believe (2nd moment) without low V;JQ or high VA/Q areas. Note also the that increasing inspired Po2 will either release hypoxic absence of shunt. Each individual data point represenu a particular pulmonary vasoconstriction, or cause absorption atelec amount of bloodflow or alveolar ventilation. Total cardiac bloodflow corresponds to the sum of all bloodflow poinu just as total alveolar ~i~. or both [21, 22), and thereby change the very ventilation is the sum of all ventilation points. V A/Q distribution that is being measured. Second, by administering all six gases simultaneously, an internally consistent picture is obtained. Using techniques of While the original technique requires estimates of inert sequential administration of different oxygen concentra gas concentrations in three sites: mixed expired gas, tions, changes could occur over time that would arterial blood, and pulmonary arterial blood, it is confound the issues. Most importantly, the results are possible to use the method without having to directly not intrinsically affected by the same variables that will sample from all three sites. Thus, an alternative method alter venous admixture or physiol.ogi~al dead space para employs direct measurement of mixed expired gas and meters even if the underlying VA/Q distribution is not arterial blood inert gas concentrations, but utilizes meas altered (factors such as changes in cardiac output or total ured values of cardiac output to calculate by mass bal alveolar ventilation). Thus, under any given set of ance the corresponding mixed venous concentrations. This conditions, the appropriate VA/Q distribution is obtained. has been shown to give results as good as the approach Finally, while two- and three-compartment approaches in which pulmonary arterial blood is directly sampled. to quantifying VA/Q inequality are at best poorly able to The method can be used in situations where an indwell separat~ lo~ VA/Q areas (poorly ventilated Jung) from ing pulmonary catheter is not indicated or may not be shunt (VA/Q=O) on the one hand, and high 'Y A/Q areas placed. The use of a venous line to inject indocyanine (poorly perfused lung) from dead space (VA/Q=infinity) green for measurement of cardiac output has also been on the other, the MIGET is, by its nature, better able to successfully applied in several clinical settings [17, 18]. achieve this function than any other former approach. 474 R. RODRIGUEZ·ROJSIN, P.D. WAGNER The reason for this is straightforward and resides in the It otherwise appears to be an undetectable component of utilization of several gases of different but appropriate pulmonary gas exchange, both in health and disease. solubilities. Consequently, the quantitative relationships between the gases is an important component that allows the resolution amongst units of different VA/Q ratios. Clinical examples of the information Nevertheless, some aspects of pulmonary gas exchange available from MIGET cannot be assessed by the MIGET. For instance, the inert gas me.th<>Q cannot separate the effects of parallel versus In the text below we will review three different series V A/Q heterogeneity if series inequality is present examples of clinical respiratory problems, where the In other words, the inert gas method can be analysed on complex interactions of intrapulmonary and extrapulmon the basis of a parallel model even if series inequality is ary factors regulating 0 2 and C02 are specifically present, but that series inequality is not identifiable by assessed on the basis of the results afforded by the this approach [23]. Re-inspiration of gas from series MIGET. In the frrst condition, we will discuss the issue dead space may result in the underestimation of meas of the gas exchange response to bronchodilators in ured dead space and a slight alteration in the V A/Q bronchial asthma, a disease affecting primarily the air distribution in the middle range [24]. Acetone is the ways. In the second example, we will deal with a pulmo most soluble of the gases used, and an accurate measure nary vascular disease, acute pulmonary embolism, and ment of the excretion of acetone (by avoiding condensa the interplay of the different mechanisms governing tion of water vapour through the expired circuit and also abnormal gas exchange. Finally, in the last condition, the the contaminant effects of some brands of heparin on major determinants of hypoxaemia that develops during blood [25]) is necessary to distinguish between dead space haemodialysis, a respiratory disorder of extrapulmonary ~en~ilation from ventilation to lung units with high origin, are explored by means of the results given by the VA/Q ratios. MIGET. The inert gas elimination technique can assist not only in the delineation of VA/Q mismatching and shunt providing a quantitative estimate of these phenomena, Bronchial asthma and gas exchange response to but is also a useful tool in addressing the potential bronchodilators presence of alveolar to endcapillary diffusion limitation for oxygen. As shown previously (and assumed in In 1978, WAGNER et al. [3) measured VA/Q inequality equation 1), inert gases equilibrate across the lung by means of the MIGET in patients with asymptomatic extremely rapidly [26]. Thus, within the first few asthma. All except one showed clearly bimodal YA/Q hundredths of a second of the red cells' exposure to distributions but no sh\)nt .(fig. 3). The presence of a alveolar gas in its transit through the lung capillary, there bimodal distribution of V A/Q ratios was compatible with is partial pressure equilibration of all inert gases between the presumed pathophysiology of gas exchange abnor alveolar gas and blood. Oxygen, however, requires an malities in bronchial asthma, namely airway narrow order of magnitude longer to achieve essential equilibra ing caused by generalized bronchoconstriction coupled tion, approximately 0.2-0.3 sin normal conditions [27]. with mucus plugging and/or bronchial wall oedema. Such Under such conditions, neither oxygen nor the inert gases airways obstruction produces low VA/Q areas that are are diffusion-limited in the lung because capillary transit well perfused but not sufficiently ventilated. The absence time is of the order of 0.75 s. However, in the presence of shunt was thought to be due to the efficiency of of pathology and/or during exercise, particularly under collateral ventilation which is considered an important hypoxic conditions, oxygen may not reach equilibrium mechanism maintaining alveolar ventilation beyond between alveolar gas and endcapillary blood in a single obstructe9 ajrways. A finding of considerable interest transit. This can therefore generate alveolar-arterial Po was that VA/Q mismatch can be present despite a nearly • differences and hypoxaemia. Because the inert gase~ normal Pao2 and Paco2 This finding, as discussed equilibrate so much more quickly than oxygen, their above, exemplifies how the degree of arterial hypoxae exchange is still not diffusion limited. It is possible mia caused by a given amount of VA/Q abnormality depends considerably on other factors, such as cardiac therefore to quantitatively compute the arterial Po2 that • wo~;~ld l'e expected (predicted) to occur from the amount output, which in turn increases the PVo2 The increased of V A/Q mismatch detected by the inert gas technique. cardiac output was related to the anxiety of the experi mental study, and possibly to some residual bronchodi When this is done, it can be compared to the actual Pao2 measured in the subject. If these agree statistically, one lating effects from earlier regular treatment. can conclude that diffusion limitation is not occurring Another conspicuous finding of this study was the for oxygen. Under certain conditions, however, diffusion presence of a transient deterioration in both VA/Q limitation for oxygen may be present, and then one would relationships and Pao2, together with a return to normal expect more hypoxaemia than can be explained by the in maximal airflow rates (FEV1 and forced expiratory inert gas elimination data. This has been found to occur flow between 25- 75% of forced vital capacity in normal subjects heavily exercising, especially when (FEF2..S-75J) after nebulizcd isoproterenol (fig. 3). It was breathing hypoxic gas mixtures [28, 29] and also in concluded that the I!laj9r factor contributing to this patients with restrictive lung diseases both at rest [30] further worsening of VA/Q relationships was an increase and during exercise breathing room air [2, 30]. in cardiac output. Release of hypoxic pulmonary VENTrLATION-PERFUSION INEQUALITY 475 1.4 I 4 ~ i:: D.M. asthma prior to bronchodilator e D.M. asthma 5 mln post bronchodilator e.... 1.2 .... I 2 1.0 10 it Pao = 81 mmHg 3: Pao, = 70 mmHg S2 1 S2 l; 0.8 08 0 ~0 0 :a :a 0.6 ... 06 ... 0 0 ~ ~ 0 4 S2 0.4 S2 ~ ~ ;:; 0.2 E 02 c Cll Cl) > 1 No Sllv111 > 0 0 0 0.01 0.1 LO 100 1000 0 001 0 1 10 100 100.0 Ventllatlon-perfuslon ratio Ventllatlon-perfuslon ratio ~ 1.4 i:: 14 e D.M. asthma 10 mln post bronchodilator e D.M. asthma 20 mln post bronchodilator .... 1.2 .... I 2 it 1.0 Pao, =82 mmHg it 10 Pao, = 84 mmHg S2 S2 ~ 0.8 ~ 08 Vt nflltJ!lon S2 0 / .Q 816H Fl#lf :a ... 0.-6 ... 06 0 "'-.. 0 c c 0.4 0 04 S2 ;:; ~ 2! E 0.2 E 0.2 Cl) ,. SM/11 Cl) > I > /ill Sllvllf 0 0 0 0.01 0.1 tO 10.0 100.0 0 0.01 0.1 tO 10.0 100.0 Ventllatlon-pertuslon ratio Ventllatlon-perfuslon ratio Fig. 3. - Representative s~uence of ventilation-perfusion (v A/Q) distributions in an asthmatic patient taken from the study of WAONI!a et al. [3}. At baseline, notice the bimodal bloodflow distribution (averaging approximately 20% of the cardiac output) but no shunt in the face of a normal anerial oxygen tension (Pao ). five mlnutes after isoproterenol, the low VA/Q areas contained twice as much pulmonary perfusion as before the drug was given and Pao feif 11 mmHg. By 10 and 20 min after bronchodilator, the distributions of ventilation and bloodflow were not essentially 1 different from pre-bronchodilatation results and Pao values were within normal limits. (Reprinted with permission of the American Review of 1 Respiratory Disease). vasoconstriction was not though.t as. important because mechanisms could singly, or in combination, account for the fraction of bloodflow to low VA/Q ratio units did not this. First, due to an increase in total perfusion and/or increase after 100% 0 2 breathing in four of these selective reversal of hypoxic vasoconstriction, existing patients, as it did after isoproterenol. Unfortunately, it low VA/Q areas could have their perfusion increased. is not possible, from these data, to differentiate the The second possibility is that new areas of low VA/Q separate roles of active vasodilatation reducing the pul ratio could have been created simply by overperfusion monary vascular tone and increased cardiac output in without reduction in xen~ilation . The third possibility is producing augmented fractional bloodflow to low '-IA/Q that new areas of low V A/Q ratio could be created through ratio areas. Indeed, data from CaRTE and YoUNG [31] and reduction in ventilation of areas in the lung. From the also from our own group [32, 33] have shown subse strict mathematical standpoint, data such as those illus quently that patients with asthma, particularly those with trated in figure 3 cannot distinguish amongst these three the most severe exacerbation of the disease, elicit a possibilities. Common sense, however, would suggest that considerable hypoxic vascular response breathing 100% when a bronchodilator is given, and at the same time 0~ flow rates at the mouth are shown to improve, local The increased perfusion of low VA/Q areas associated reduction in ventilation is an unlikely candidate for the with bronchodilator inhalation at 5 min in figure 3 raises changes.shQwn in the figure. In addition, to produce areas an interesting problem of interpretation. What the data of whose VA/Q ratio is of the order of 0.05 (one 20th of this panel clearly say is that a greater fraction of the normal) through overperfusion, it would require a 20- cardiac output i~ p~rfusing areas of the lung with an fold increase in local bloodflow. While mathematically abnormally low VA/Q ratio. Theoretically, three separate feasible, this would not seem physiologically likely. The 476 R. RODRIGUEZ-ROISIN, P.D. WAGNER remaining mechanism (of those listed above) is increased likewise an increase in 0 2 consumption. In. co~trast, in a per fusion of areas which are already poorly ventilated. In similar group of acute asthmatic patients V A/Q distribu the example of figure 3, this would be consistent with an tions did not deteriorate 15 min after inhaled salbutamol approximate doubling of the perfusion of these areas, (300 J.lg) while there were no changes in cardiac • which is a more reasonable possibility. Thus, while no output, 0 2 consumption, or Pao2 Yet, both forms of definitive mechanistic analysis can be given for the administration of salbutarnol significantly improved FEV1 effect of bronchodilators (or indeed of other manoeuvres to the same degree without changes ip minute ventila in other clinical conditions), an analysis of the observed tion. The differences in the effects on VA/Q abnormali changes using both common sense and logic would lead ties of salbutamol (which are summarized in table 3), one to reject what appear to be quite unreasonable administered by two different routes, may well be explanations. explained by the relevant associated cardiovascular and Since HALMAGYI and COTEs [34) draw attention to the metabolic effects (basically, increases in cardiac output finding that patients with COPD may develop arterial and 0 2 consumption, respectively) which can reasonably hypoxaemia after administration of epinephrine and be related to the greater plasma levels of salbutamol aminophylline, there have been similar observations presumably achieved during the intravenous administra following the effects of various bronchodilating drugs tion of the drug. Again, from these data there is not [35, 36]. Since then there were numerous other studies sufficient information to differentiate between an increase [37- 39) of "paradoxical hypoxaemia" occurring shortly in cardiac output provoking an increase in the amount of afler administration of bronchodilators to patients with dispersion of pulmonary bloodflow, or a reduction in asthma, in spite of improvement in airflow rates. The pulmonary vascular tone. Even more recently, our group changes noted often varied with the drug given. The [43] has also demonstrated a lack of change in VA/Q observation that the administration of bronchodilators may mismatch in patients with severe chronic asthma who alter gas exchange despite its beneficial effect on bron did show a significant improvement in FEV1 after 300 choconstriction has been one of the most provocative J.lg of inhaled salbutamol. findings in patients with bronchial asthma. Furthennore, the observation in asthmatics of lack of improvement in Table 3. - Gas exchange response to intravenous and distributions of alveolar ventilation and pulmonary per inhaled salbutamol in acute severe asthma fusion while airflow rates are returning to nonnal has raised an interesting debate. Because non-selective Intravenous Inhaled beta-adrenergic agonists (i.e., isoproterenol) have been Minute Ventilation Unchanged Unchanged thought to produce a deleterious effect on Pao2 through their well-known beta cardiovascular effects (more Cardiac Output Increased Unchanged specifically pulmonary vasodilatation secondary to ?z ~onsumption Increased Unchanged increased cardiac output), it was suggested that a more VA/Q Inequality Increased Unchanged specifically be~-agonist (for instance, salbutamol), which Pao Unchanged Unchanged may not provoke such increase in cardiac output, could 2 be associated with. a greater fall in Pao if it caused the : 2 VA/Q: ventilation-perfusion distribution; Pao2 arterial oxygen same increase in VA/Q mismatching. tension. Interestingly 1 other studies of the effects of bronchodi lators on VA/Q mismatching have yielded different Formerly, similar results were fairly well reproduced results. Young asymptomatic asthmatics challenged with by us [44] in an experimental model of canine "asthma" aerosolized antigen [40] showed that VA/Q inequality (challenged with inhaled methacholine), which was rapidly and completely reverted toward normal levels followed by aerosolized epinephrine, isoproterenol and following the administration of metaproterenol. Later, salbutamol. Following isoproterenol, since minute ven YoUNG et al. [41] reported in pati~nts in whom asthma tilation remained unchanged, increased cardiac output was induced by exercise that VA/Q mismatch improved together wjth aecreased pulmonary vascular resistance and cardiac output rose, several minutes after inhalation worsened VA/Q mismatch (intrapulmonary factor). The of salbutamol. In 1985, HEDLIN et al. [42) documented resulting deleterious effect on Pao2.!. thus, was counter that salbutamol improved the VA/Q distribution and balanced by the improvement in Pvo2 due to increased restored the blood gases to normal in all but one of a cardiac output (extrapulmonary factor). After epineph group of asthmatic children following histamine chal rine and salbutamol, a similar bronchodilator c(fect was lenge. It was not clear from their data, though, that associated with only slight deterioration in VA/Q 'VA/Q impairment was greater than that expected in this relationships. The effects of the former drug on VA/Q situation after exercise or challenge was discontinued or , distributions, Pao2 cardiac output, heart rate, and whether these changes were statistically significant, thus pulmonary vascular resistance were less and of shorter making their data difficult to interpret. Subsequently, duration than those of isoproterenol. In contrast, salbuta BALLEsTER et al. [33] have shown that, during intra mol resulted in changes similar to but more persistent venous salbutamol (4 J.lg·min·1 for 1 h) in patients with than those caused by epinephrine. These results essen ~cu~ severe asthma, there is development of further tially showed that the gas exchange responses to three VA/Q inequality along with a substantial rise in cardiac different bronchodilators in this model of broncho ; output with the outcome of an unaltered Pao2 there is constriction are closely related to the expected VENTILATION-PERFUSION INEQUALITY 477 pharmacologic effects (alpha, ~ta1 , and be~) of these Mechanisms of arterial hypoxaemia in pulmonary drugs. Moreover, they also confmned that Pao changes 2 embolism per se may not necessarily reflect changes in VA/Q inequality t?ec~use increased cardiac output may prevent The potential causes for hypoxaemia in acute pulmo increased VA/Q mismatch from resulting in a decreased nary embolism are several and different explanations have Pao2• It is of note that nebulized epinephrine (1 m g) has been postulated [49, 50]. First, VA/Q mismatching may been shown to be as effective as salbutarnol in patients be a possible factor in the arterial hypoxaemia. This can with acute asthma without associated adverse effects [45], happen by mechanical redistribution of pulmonary using conventional techniques. bloodflow a':VaY. from obstructed vascular areas. There Worsening or lack of improvement in VA/Q mismatch can also be VA/Q inequality due to resolving areas of after spirometric variables in both experimental and pulmonary infarction, atelectasis, or oedema that remain human asthma conditions have returned to normal perfused but poorly, or not ventilated at all, because of suggest that the gas exchange abnormalities are related alveolar wall stiffness and/or alveolar space filling to mucus plugging and/or bronchial wall oedema in secondary to fluid oedema or cellular debris. Another the peripheral, small airways, rather than to reversible possibility may be related to the development of bronchoconstriction in larger airways. Indeed, this humorally-induced bronchOfl10to{ changes, causing pneu hypothesis has been subsequently strengthened by RocA moconstriction, anQ h~nce V A/Q mismatching. Second, et al. [46] in patients with acute severe asthma requiring and in addition to VA/Q mismatch, there may be shunt hospitalization on admission ~d d!1Jing recovery, such ing of blood through non-ventilated alveoli caused either that spirometric variables and V A/Q inequality indicate by atelectasis owing to pulmonary surfactant loss or distinct pathophysiologic phenomena with quite different alveolar flooding by pulmonary oedema. However, fully time courses of recovery in this clinical condition. developed pulmonary oedema in pulmonary embo According to this concept, maximal airflow rates pre lism is rarely if ever seen in man. Another possibility of dominantly reflect bronchoconstriction in large, central intrapulmonary shunting is the opening of direct pulmo airways, whereas VA/Q maldistribution is more related nary arteriovenous channels provoked by high pulmo to the events that operate in the lung's "quiet zone" [47]. nary artery pressure, including the reopening of the Very recently, this hypothesis has been reinforced by foramen ovale. Thirdly, there have been suggestions of , FERRER et al. [48] who demonstrated that patients with diffusion limitation for 0 2 caused by either a reduced acute asthma who are discharged from the emergency red cell transit time or the development of interstitial room following a severe attack show that both the time oedema, that might result in an increased diffusional course and the improvement of spirometric variables and pathway for 0 2 between the lumen of the alveoli and the VA/Q mismatch data are paralleled as opposed to those red cell. who are hospitalized, whose results are not The MIGET has only recently been used in patients It becomes evident that worsening of hypoxaemia by with acute pulmonary embolism, and only a few patients bronchodilators may be a side-effect of the treatment of have been studied. D' Ar.oNZO et al. [51] was the first to patients with asthma. A small decrease in Pao2 in a patient study 2 patients with acute, massive pulmonary embo without pre-existing moderate to severe hypoxaemia is lism. Both individuals showed a considerable increase in unlikely to have any clinical significance; in contrast, a the P(A-a)o2 and also substantial VA/Q inequality, the similar decrease in a patient with pre-existing severe measurements being made one and eight days following hypoxaemia might be potentially life-threatening. How the acute episode, respectively (fig. 4). The increased ever, one can certainly not withhold bronchodilators in YA/Q inequality was due entirely to the presence of high patients with asthma because of their transient (and V A/Q areas, namely an im~re~se in the percentage of • ventilation diverted to high VA/Q regions consistent with sometimes unpredictable) deleterious effect on Pao2 In summary, taking all of these studies together, the unde~lying pulmonary vascular obstruction. Although several implications that may have clinical relevance for no low V A/Q areas were shown, both patients had a the care of p~tie~ts with asthma can be inferred. First, considerable amount of shunt (20% and 38% of cardiac considerable V A/Q inequality may be present in the face output, respectively), not evident on chest radiographs, of little arterial hypoxaemia, because of the usually as well as ventilation of totally unperfused lung units • (dead space). The increased P(A-~)o _was explained mostly elevated cardiac output preserving Pvo2 Second, 2 increased hypoxaemia may well occur following use of by the large shunt present and V A/Q mismatching had a adrenergic bronchodilators owing to an increase in the smaller contribution to the hypoxaemia of these two amount of VA/Q mismatching. Third, inhaled salbutamol patients. Diffusion limitation to 0 2 was not found to be seems to be an efficient and safe therapeutic approach, a mechanism of abnormal gas exchange in these cases. unlike other inhaled non-selective beta-adrenergic agents, Two years later, MANIER et al. [49] pointed out that the because of the lack of deleterious effect on 'V A/Q mis decrease in Pvo2 constitutes the major determinant of match for the same degree of bronchodilatation. Cardiac hypoxaemia in acute pulmonary embolism, and is output changes, therefore, become important in govern enhanced by the associated VA/Q inequality. The latter ing overall gas exchange abnormalities in bronchial was noted to be a moderate amount of bloodflow asthma, particularly during the intravenous administra distributed in low VA/Q ratio units, and also shunt (mean tion of the most beta-selective adrenergic bronchodilating 9.% Qf cardiac output). The amount of ventilation in high drugs or after any of the inhaled non-selective ones. VA/Q areas was considerable (mean 54%). 478 R. RODRIGUEZ-ROISIN, P.D. WAGNER Interestingly, 13% of the increased P(A-a)o1 was (range 3-17% of cardiac output) were shown, along with explained by an 0 2 diffusional limitation componenL radiological evidence of atelectasis or alveolar filling, However, a major breakthrough in our underslanding of and these shunts accounted for most of the hypoxaemia. the underlying gas exchange abnormalities in patients It would be interesting, however, to conflrm these data with pulmonary embolism has come from HUET et al. by assessing sequentially the evolution of 'i.IAJQ abnor (50] who documented 7 patients with severe, acute malities in a larger series of patients with pulmonary pulmonary embolism. These authors emphasized the role embolism. In all the patients, ventilation to unperfused of a low PVo+ in the presence of either VA/lJ. mismatch areas (dead space) was markedly increased (mean 52% or shunt whtch contributed to hypoxaemia in these of ventilation). patients. Sim~lar\y, as in the report of D'ALoNZO et al. In summary, based qn ·~ese few studies it seems that, (51]. both VA/Q inequality and shunt were evident. first, both shunt and VA/Q mismatching are major con sequences of severe, acute pulmonary embolism. 1.0 Ventilation-perfusion inequality includes the presence of t t 0.11 Shunt 20% Dead both low and high VA/Q areas. Second, shunt seems to apec:• be more important in the subacute stages of the disease 42% o.e o.u. Vr:-13.2 #mlrr1 whereas part of the VA/Q mismatch, more specifically 0.1 C0:3.7/·mJn·• bloodflow perfusing lung units with low VA/Q ratios, 0.6 plays a major role during the early phases as important ':' c o.s factors modulating hypoxaemia. Third, mixed venous Po1 ·e i_s a.major factor, in the presence of either shunt and ... 0.4 • V A/Q mismatch, contributing to the low values of Pao2 ...... 0 .3 Fourth, diffusion limitation for 0 2 is hardly altered such l 0.2 that there is little consistent evidence to support a 0 . 1 significant contribution of this factor for determining ~ 0 hypoxaemia in pulmonary embolism. ;:: "C 0 0 .01 0 .1 10 10 100 0 :D Mechanisms of abnormal gas exchange in haemodialysis "C c 1.0 CO t t o.a Shunt 39% Dead Diffusion impairment [52, 53], hypopnoea (alveolar space hypoventilation without associated hypercapnia) accom o.e 69% B.O. panied by a decrease in the respiratory exchange ratio Vr.:17.0 fmln·• I 0 .7 c C0:4.2 fmln'' (RQ) (54-59], and VA/{J inequality [60-62] have all been 0 += O.G implicated as mechanisms underlying the development .!l! += o.s of arterial hypoxaemia during haemodialysis. However, c: it is essentially impossible to agree on which factors are ~ 0 .4 most important by using conventional gas exchange 0.3 techniques such that the question remained unanswered 0 .2 for some time. Accordingly, we used the MIGET (63] in 0 1 a group of patients with chronic renal failure before, during, and after regular haemodialysis to more critically assess the mechanisms of hypoxaemia by directly meas 0 01 0 1 1.0 10 100 uring the VA/Q distribution. As did others [64]. we found Ventilatlon-perfusion ratio a significant fall in Pao2 during the procedure. Concomi· Fig. 4. - Ventilation·perfusion distributions recovered by D'ALONZO ~~ lantly, there were significant decreases in minute venti al. [50] in 2 patients (O,U. and B. D.) with acut~ pulmonary embolism lation, alveolar ventilation, col elimination throughout showing considerable ventilation-perfusion (VA/Q) ineqqality and the longs, and RQ since pulmonary 0 2 consumption substantial amounts of shunt and of dead space. Nots: th,at VA/Q mis remained unchanged. As a result, al veolar Po fell, match was characterized by the presence of high VA/Q areas (i . ~ .• 1 , increased ventilation in high V A/Q ratio units} (see te;>tt for funher essentially by as much as Pao2 such that P(A-a)o2 was explanation). CO: cardiac output; Ve: minute ventilation rate. unchanged. Cardiac output also dropped significantly and (Reprinted with permission of the American Review of Respiratory there was a significant reduction in the percentage of Disease. blooc)flow to nonventilated (shunt) or poorly ventilated HO\yev~r. the imporlant observation was the time courses low V A/Q ratio units during haemodialysis (which iiJon~ ) of VA/Q abnormalities and both the clinical and radio would have raised Pao2 (fig. 5). Distributions of VA/Q graphic findings. In this regard, VA/Q mismatching ratios were ms>derately abnormal b.efo~e dialysis, includ accounted for most of the hypoxaemia in the 2 patients ing both low V A/Q areas and high V A/Q areas (almost all assessed early after embolism (i.e., less than 48 h 9f ~is was due to dead space). Therefore, if anything, after the acute episode), whose chest X-rays were nor· VA/Q mismatching improved as Pao2 decreased. The most mal. In the other 5, in contrast, whose assessment was likely cause of this finding is a decrease in extravascular done between 84-216 h after the occurrence of pulmo lung water due io haemodialysis (65]. although it may nary embolism, moderate to large amounts of shunt well also be the result of the simultaneous fall in cardiac VENTll-ATION-PERFUSION INEQUALITY 479 120 mln 11.5 kPa BaMIIne 10.7 kPa 60 mln 075 Pao, 10.5 kPa 4.87. lmln·• 3.64 lmln·• 3.14 fmln·• 4.46 lmln'' 3.71 lmln·• "'•Or 3.25 lmln·• 0.50 0.50 050 - 1: E o2s 025 .... 025 e..... - • - .L 0 'I. oo "' 01 1.0 10.0 ICOO a 001 01 1.0 100 1000 ;:~ "C 0 0 .Q "C 210 mln Aher c ~eo, 10.8 kPa Pao, 12.1 kPa CV 075 VA 4.85 lmln·• v. 5.81 lmln·• Or 3.18 lmln·• 1()()1 0. 4.47 lmlli• 0.75 0.50 050 02, 0.25 1.7'~ 0 • 0 001 0 .1 lO 100 1000 10.0 1000 Fig. 5. - Representative serial ventilation-perfusiol} (v_,./Q) distributions of one patient with chronic renal failure in the study of RoMALDINI et al. [59]. There is a progressive improvement in low VA/Q areas (from 11.4% of bloodflow before haemodialysis to 8.6, 5.5, and 1.7% at 60, 120, ) A) apd 210 min during haemodialysis, respectively); arterial oxygen tensiol} (Pao2 also decreased as did alveolar ventilation (V and cardiac output (QT). After haemodialysis, all but low VA/Q areas (with only 4.3% of Qr) of the variables returned toward baseline values. (Reprinted with permission of the American Review of Respiratory Disease output The finding that perfusion of low VA/Q areas did red cell transit in the rest of opened capillaries, was increase again as soon as dialysis was discontinued, unlikely since we found that patients dialyzed with poly suggests that some of the VA/Q improvement was related acrylonitrile, as opposed to those done with cuprophan, to the transiently decreased cardiac output. Because this membranes decreased their Pao2 just as much despite the increase was not toward pre-dialysis values, our data absence of leucopenia. Alternatively, obstruction of therefore suggest an effect of both decreased cardiac capillaries by leucocyte microaggregation should increase, output and possible t:edqction in extravascular lung water if anything, the ventilation in high VA/\). areas, a finding acting to improve V A/Q mismatching. Similar results, not shown in our study [63]. Finally, the percentage also using the MIGET, have been shown in a canine reduction in carbon monoxide diffusing capacity (DLCO) model of dialysis [66]. shown previously by other investigators [53, 61] would • The question arises as to whether the relationship not be sufficient to explain the observed falls in Pao2 between cardiac output and nonventilated (shunt) or poorly ventilated areas is just another manifestation of Summary and conclusions the well-known, but poorly understood, positive associa tion between shunt and cardiac output [67, 68] or whether This review has shown the extent of information haemodialysis actually reduces VA/Q maldistribution. We available from the MIGET, such that 1) an estimate of suggest that both of these mechanisms coexist and oper the pattern of ventilation and perfusion is obtained from ate to explain our findings. which; 2) numerical parameters describing the quan The hypothesis of a role for 0 2 diffusion limitation as titative abnormalities can be readily calculated; 3) the an additional mechanism for arterial hypoxaemia seems presence of diffusion limitation of pulmonary oxygen less likely. It was demonstrated that the combined effect transport can be inferred; and 4) the alveolar-arterial Po2 ~f ~educed alveolar ventilation, cardiac output, and difference can be partitioned into components due to VA/Q mismatch was sufficient to account for all of the shunt, VA/Q inequality and alveolar-endcapillary observed hypoxaemia. Another argument, leucocyte diffusion limitation for oxygen; finally, 5) changes over microaggregation in the pulmonary capillary network time in arterial oxygenation can be apportioned into during haemodialysis [53, 60-62], with subsequent rapid intrapulmonary and extrapulmonary factors to improve 480 R. RODRIGUEZ-ROISTN, P.D. WAGNER our understanding of the physiological basis of these gas method for estimating VA/Q maldistribution. Respir changes, especially critical when many of the deter Physiol, 1982, 49, 293-313. minants of Pao may be changing at the same time. 13. Kapitan KS, Wagner PD.- Linear programming analysis In spite of a lew limitations [23-25] the multiple inert of VA/Q distributions: limits on central moments. J Appl Physiol, 1986, 60, 1772-1781. gas elimination technique has proven to be a useful 14. Wagner PD. Naumann PF, Laravuso RB.- Simultaneous method in the analysis of pulmonary gas exchange. Yet, measurement of eight foreign gases in blood by gas it is sometimes regarded as a formidable research tool chromatography. J Appl Physio/, 1974, 36, 600-605. not generally well-suited to clinical application. At the 15. Wagner PD. L6pez FA.- Gas chromatography teclmiques present time, the technique is limited by the time elapsed in respiratory physiology. In: Techniques in the Life Sciences. for the gas extraction to process the blood samples and, A.B. Otis ed., Elsevier Scientific Publishers Ireland Ltd, County also, the time for manual conversion of chromatographic Clare, Ireland, 1984, pp. 403/1-403/24. data to digital data for the computer. This may well be 16. Wagner PD. Hedenstiema G, Bylin G, Lagcrstrand L. - overcome in the future by using new technology, such as Reproducibility of the multiple inert gas elimination technique. rapid extraction from blood and continuous on-line J Appl Physiol, 1987, 62, 1740-1746. 17. Wagner PD, Laravuso RB, Uhl RR, West JB. - measurements [69) with direct links between chromato Continuous distributions of ventilation-perfusion ratios in nor graph and computer. Meanwhile, studies using the mal subjects breathing air and 100% 0 2• J Clin Invest, 1974, peripheral venous approach, such as those recently done 54, 54-Q8. in patients with different clinical forms of asthma [20, 18. Gale GE, Torre-Bueno J, Moon RE, Saltzman HA. Wagner 46, 48), have clearly demonstrated that the conclusion of PD. - Ventilation-perfusion inequality in normal humans clinical nonsuitability need not be true. The use of inert during exercise. 1 Appl Physiol, 1985, 58, 978-988. gases to characterize gas exchange in the lung has thus 19. Wagner PD. Smith CM, Davies NJH, McEvoy RD. Gale provided us with a new level of understanding of respi GE. - Estimation of ventilation-perfusion inequality by inert ratory pathophysiology in disease. The method can be gas elimination without arterial sampling. J Appl Physiol, 1985, used to follow clinical and therapeutic interventions 59, 376-383. 20. Wagner PD, Hedenstiema G, Bylin G. - Ventilation and as a research tool in pulmonary medicine. Possibili perfusion inequality in chronic asthma. Am Rev Respir Dis, ties for future research in the field of gas exchange 1987, 136, 60.5-612. pathophysiology with this technique continue to be 21. Briscoe W A. Cree EM, Filler J, Houssay HEJ, Coumand exciting. A. - Lung volume, alveolar ventilation and perfusion inter relationships in chronic pulmonary emphysema. J Appl Physiol 1960, 15, 785-795. References 22. Dantzker DR. Wagner PD. West IB.- Instability of lung units with low VA/Q ratios during ol breathing. J Appl 1. West JB. - Causes of carbon dioxide retention in lung Physiol, 1975, 38, 886-895. disease. N Engl J Med, 1971, 284, 1232-1236. 23. Wagner PD. Evans JW.- Conditions for equivalence of 2. Wagner PD. - Ventilation-perfusion inequality and gas gas exchange in series and parallel models of the lung. Respir exchange during exercise in lung disease. In: Muscular Exer Physiol, 1977, 31, 117-138. cise and the lung. J.A. Dempsey, C.E. Reed eds, University of 24. Fortune IB, Wagner PD.- Effects of common dead space Wisconsim Press, Madison, 1977, pp. 345-356. on inert gas exchange in mathematical models of the lung. J 3. Wagner PD, Dantzker DR. Iacovoni VE, Tomlin WC, West Appl Physiol, 1979, 47, 896-906. IB. - Ventilation-perfusion inequality in asymptomatic asthma. 25. Powell FL. L6pez FA, Wagner PD.- Effects of acetone Am Rev Respir Dis, 1978, ll8, 511-524. in heparin on the multiple inert gas elimination technique. J 4. West IB. - Ventilation-perfusion inequality and overall Appl Physiol, 1985, 58, 1143-1157. gas exchange in computer models of the lung. Respir Physiol, 26. Forster RE. - Exchange of gases between alveolar air and 1969, 7, 88-110. pulmonary capillary blood: pulmonary diffusing capacity. 5. Wagner PD. - Interpretation of arterial blood gases. Chest, Physiol Rev, 1957, 37, 391-453. 1980, 77, 131-132. 27. Wagner PD.- Diffusion and chemical reaction in pulmo 6. Fenn WO, Rahn H. Otis AB. - A theoretical analysis of nary gas exchange. Physiol Rev, 1967, 57, 257-312. the composition of alveolar air at altitude. Am J Physiol, 1946, 28. Torre-Bueno J, Wagner PD, Sa1tzman GR, Gale GE, Moon 146, 637-653. RE. - Diffusion limitation in normal humans during exercise at 7. Riley RL, Coumand A. - "Ideal" alveolar air and the sea level and simulated altitude. J Appl Physiol, 1985, 58, analysis of ventilation-perfusion relationships in the lungs. J 989-995. Appl Physiol, 1949, 1, 825-847. 29. Wagner PD, Gale GE, Moon RE, Torre-Bueno J, Saltzman 8. Kety SS. -The theory and applications of the exchange of HA. - Pulmonary gas exchange in humans exercising at inert gas at the lungs and tissues. Pharmacol Rev, 1951. 3. sea level and simulated exercise. J Appl Physiol, 1986, 61, 1-41. 260-270. 9. Farhi LE. - Elimination of inert gas by the lung. 30. Agustf AGN, Roca J, Rodriguez-Roisin R, Gea J, Xaubet Respir Physio/, 1961, 3, 1-11. A, Wagner PD.- Role of 0 1 diffusion limitation in idiopathic 10. Wagner PD, Saltzman HA, West JB.- Measurements of pulmonary fibrosis (abstract). Am Rev Respir Dis, 1987, 135, continuous distributions of ventilation-perfusion ratios: theory. A307. J Appl Physiol, 1974, 36, 588-589. . . 31. Corte P, Young IH.- Ventilation-perfusion relationships 11. Evans JW, Wagner PD. - Limits on VA/Q distributions in symptomatic asthma. Response to oxygen and clemastine. from analysis of experimental inert gas elimination. J Appl Chest, 1985, 88, 167-175. Physiol, 1977, 42, 889-898. 32. Rodriguez-Roisin R, Ballester E, Roca J, Torres A, 12. Ratner ER, Wagner PD. - Resolution of the multiple inert Wagncr PD. - Mechanisms of hypoxaemia in patients with VENTILATION-PERFUSION INEQUALITY 481 status asthmaticus requiring mechanical ventilation. Am Rev 53. Mahajan SK, Gardiner WH, De Tar B, et al. - Respir Dis, 1989, 139, 732- 739. Relationship between pulmonary function and hemodialysis 33. Ballester E, Reyes J, Roca J, Guitart R, Wagncr PD, induced leukopenia. Trans Am Soc Artif Intern Organs, 1978, Rodriguez-Roisin R. - Ventilation-perfusion mismatching in 23, 411-415. acute severe asthma. Effects of salbutamol and 100% oxygen. 54. Aurigemma NM, Feldman NT, Gonlieb M, Ingram RH Jr, Thorax, 1989, 44, 258- 267. Lazsrus JM, Lowrie EG. - Arterial oxygenation during 34. Halmagyi DF, Cotes JE. - Reduction in systemic blood hemodialysis. N Eng/ J Med, 1977, 297, 871- 873. oxygen as a result of procedures affecting the pulmonary cir 55. Sherlock JE, Ledwith J, Letteri J. - Hypoventilation and culation in patients with chronic pulmonary disease. Clin Sci, hypoxaemia during hemodialysis: reflex response to removal 1959, 18, 475-489. of C02 across the dialyzer. Trans Am Soc Artif Intern Organs, 35. Read J, Pain MCF, Tai E. - Effect of aminophylline 1977, 23, 406-410. given intravenously on arterial oxygen tension in emphysema 56. Martin L. - Hypoventilation without elevated carbon and asthma. Aust Ann Med, 1965, 14, 334. dioxide tension. Chest, 1980, 77, 6. 36. Tai E, Read J.- Response of blood gas tensions to amino 57. Dolan MJ, Whipp BJ, Davidson WD, Weitzman RE, phylline and isoprenaline in patients with asthma. Thorax, 1967, Wasserman K. - Hypopnea associated with acetate hemodia 22, 543- 549. lysis: carbon dioxide-flow dependent ventilation. N Eng/J Med, 37. Knudson RI, Constantine HP. - An effect of isoproterenol 1981, 305, 72- 75. on ventilation-perfusion in asthmatic versus normal subjects. J 58. Patterson RW, Nissenson AR, Miller J, Smith RT, Narino Appl Physio/, 1967, 22, 402-406. RG, Sullivan SF.- Hypoxaemia and pulmonary gas exchange 38. Legge JS, Gaddie J, Palmer KNV.- Comparison of two during hemodialysis. J App/ Physiol, 1981, 50, 259-264. oral bet~ adrenergic stimulant drugs in bronchial asthma. Br 59. Romaldini H, Stabile C, Faro S, Lopes Dos Santos M, Med J, 1971, 1, 637-639. Tamos OL, Ribeiro Ratto 0. - Pulmonary ventilation during 39. Rees HA, Millar JS, Donald KW.- Adrenaline in bron hemodialysis. Nephron, 1982, 32, 131- 134. chial asthma. Lancet, 1967, 2, 1164-1167. 60. Craddock PR, Fehr J, Dalmasso AP, Brigham XL, Jacob 40. Wagner PD, Ramsdell JW, Incaudo GA, Rubinfeld AR, HS. - Hemodialysis leucopenia: pulmonary vascular leucosta Young IH. - Gas exchange following bronchial challenge with sis resulting from complement activation by dialyzer cellophane antigen in patients with extrinsic asthma (abstract). Am Rev membranes. J Clin Invest, 1977, 59, 879--888. Respir Dis, 1978, 117, 409. 61. Craddock PR, Fehr J, Brigham KL, Kronenberg RS, Jacob 41. Young IH, Corte P, Schoeffel RE. - Pattern and time HS. - Complement and leukocyte-mcdiated pulmonary dys course of ventilation-perfusion inequality in exercise-induced function in hemodialysis. N Engl J Med, 1977, 296, 769- 774. asthma. Am Rev Respir Dis, 1982, 125, 304-311. 62. Carlon GC, Campfield PB, Goldiner PL, Tumbull AD. - 42. Hedlin G, Freyschuss U, Hcdenstiema G. - Histamine Hypoxaemia during hemodialysis. Crit Care Med, 1979, 7, induced asthma in children: effects on the ventilation-perfusion 497-499. relationship. Clin Physiol, 1985, 5, 19-34. 63. Romaldini H, Rodriguez-Roisin R, L6pez FA, Ziegler TW, 43. Ballester E, Roca J, Rarnis Ll, Wagner PD, Rodriguez Bencowitz HZ, Wagner PD. - The mechanisms of arterial Roisin R. - Pulmonary gas exchange in severe chronic asthma. hypoxaemia during hemodialysis. Am Rev Respir Dis, 1984, Response to 100% oxygen and salbutamol. Am Rev Respir 129, 780-784. Dis, 1990, 141, (in press). 64. De Backer W A, Heyrman RM; Wittesaele WM, Van 44. Rodriguez-Roisin R, Bencowitz HZ, Ziegler MG, Wagner Waeleghem JP, Vermeire PA, De Broe ME.- Ventilation and PD. - Gas exchange response to bronchodilators following breathing patterns during hemodialysis-induced carbon dioxide methacholine challenge in dogs. Am Rev Respir Dis, 1984, 130, unloading. Am Rev Respir Dis, 1987, 136, 406-410. 617- 626. 65. Zidulka A, Despas PJ, Milic-Emili J, Anthonisen NR. - 45. Coupe MO, Guly U, Brown E, Barnes PD. - Nebulised Pulmonary function with acute loss of excess lung water by adrenaline in acute severe asthma: comparison with salbutamol. hemodialysis. Am J Med, 1973, 55, 134-141. Eur J Respir Dis, 1987, 71, 227-232. 66. Ralph DD, Ott SM, Sherrard DJ, Hlastala MP. - Inert 46. Roca J, Ramis Ll, Rodriguez-Roisin R, Ballester E, gas analysis of ventilation-perfusion matching during ?yfo~tserrat JM, Wagner PD. - Serial relationships between hemodialysis. J C/in Invest, 1984, 73, 1385- 1391. VA/Q inequality and spirometry in acute severe asthma requir 67. Breen PH, Schumacker PT, Hedenstiema G, Ali J, Wagner ing hospitalization. Am Rev Respir Dis, 1988, 137, 605- 612. PD, Wood LDH. - How does increased cardiac output increase 47. Mead J. - The lung's "quiet zone". N Eng/ J Med, 1970, shunt in pulmonary edema?. J App/ Physiol, 1982, 53, 23, 1318-1319. 1487-1495. 48. Ferrer A, Rodriguez-Roisin R, Roca J, L6pez FA, Wagner 68. Lynch JP, Mhyre JG, Dantzker DR. - Influence of car PD, Agusti'-Vidal A. - Ventilation-pcrfusion mismatching in diac output on intrapulmonary shunt. J Appl Physio/, 1979, 46, acute severe asthma in the emergency room (abstract). Am Rev 315-321. Respir Dis, 1988, 137, A408. 69. Gale GE. - In: Distributions of blood flow, ventilation. 49. Manier G, Castaing Y, Guenard H. - Determinants of tissue and gas volume in normal and edematous lungs. hypoxaemia during the acute phase of pulmonary embolism in Doctoral Dissertation. University of California, San Diego, 1987. humans. Am Rev Respir Dis, 1985, 132, 332-338. 50. Huet Y, Lemaire F, Brun-Buisson Ch, et al. - Signification c/inique des inegalites de ventilation-perfusion Hypoxaemia in acute pulmonary embolism. Chest, 1985, 88, determinees par elimination des gaz inertes. R. Rodriguez 829- 836. Roisin, P. Wagner. 51. D' Alonzo GE, Bower JS, DeHart P, Dantzker DR. - The RESUME: La premiere partie de cette revue concerne les mechanisms of abnormal gas exchange in acute massive pul mecanismes de base et les facteurs determinants de l'hypoxemie monary embolism. Am Rev Respir Dis, 1983, 128, 170-172. et de l'hypercapnie, ainsi que les differentes approches utilisees 52. Bischel MD, Scale BG, Mohler JG. - Evidence for en practique clinique et dans la recherche clinique, pour pulmonary microembolization during hemodialysis. Chest, 1975, determiner la presence d'une mauvaise congruence de la 67, 335-337. ventilation et de la perfusion, celle de shunts et des limitations 482 R. RODRIGUBZ-ROISIN, P.D. WAGNER de la diffusion pour I'oxygene, en particulier par la m6thode reponse des echanges gazeux apres administration de d'elimination de gaz inertes multiples (MIGET), en m6decine bronchodilatateurs dans l'asthrne bronchique - une maladie des pulmonaire. La seconde partie passe en revue trois maladies voies aeriennes, Jes determinants majeurs des anomalies des respiratoires differentes, oil !'interrelation complexe entre les echanges gazeux dans l'ernbolie pulmonaire aigul - une maladie facteurs intrapulmonaires et extrapulmonaires reglant l'oxygene vasculaire pulrnonaire, et enfm les memes anomalies au peut etre interpretee essentiellement par Jes resultats obtenus au cours de l'hemodialyse - une entire respiratoire d'origine cours de la demiere decennie par la technique MIGET. Sont extrapulrnonaire. explores successivement grace a la methode des gaz inertes, la Eur Respir J., 1990, 3, 469-482.