Thorax 1992;47:971-978 The pulmonary physician and critical care * 6 Series editor: T WEvans Oxygen transport: the relation between oxygen delivery and consumption R M Leach, D F Treacher The unifying concept of oxygen transport has until recently been neglected by both car- diologists and respiratory physicians. With the increasing part played by both disciplines in the care of the critically ill, however, this attitude is changing. The primary function of the heart and lungs is to generate a flow of oxygenated blood to the tissues to sustain aerobic metabolism. The main requirements of this system are that it should be energy efficient, so that unnecessary cardiorespiratory work is avoided, and that it should be sensitive to the fluctuating demands of cellullar metabolism. Secondly, metabolic demand and distribution should be matched regionally at rest, during exercise, and in different disease states. Thirdly, oxygen should be able to pass efficien- tly across the extravascular tissue matrix. The mechanisms controlling oxygen distribution are incompletely understood, but are almost certainly important in determining clinical out- come in the critically ill patient.' Although the relation between oxygen delivery (Do2) and consumption (Vo2) has not been clearly established, these variables are often measured to define a population of critically ill patients in whom Vo2 is limited by Do2, the state of so called "pathological supply dependency."2 During recent years many of the publications on critical care, and indeed practice in leading intensive care units, have emphasised the importance of raising Do2 to "supranormal" levels in an attempt to satisfy the increased metabolic demands of these patients. This practice has been justified by the observation that increased DO2 improves oxygen debt and outcome in postoperative surgical patients requiring intensive care.3 No one would dispute that restoring blood volume to improve DO2 in the severely hypovolaemic patient must be beneficial. Controlled trials, however, examining the influence of such strategies on clinical outcome in patients with more complex conditions, suffering from sepsis, cardiovascular collapse, and hypoxic hypox- aemia, have produced conflicting data. Perhaps the concept of global oxygen delivery has failed to emphasise the importance of the regional distribution of blood flow, particularly to the splanchnic and renal vascular beds, which may be more important in determining clinical outcome.' This article reviews current ideas about the relation between DO2 and Vo2, the physio- logical mechanisms controlling regional lo2, and their relevance to critical illness and currently used therapeutic interventions. Physiology of oxygen transport GLOBAL OXYGEN DELIVERY Under normal resting conditions global DO2 (defined as the product of cardiac output and arterial oxygen content) is more than adequate to meet tissue oxygen demands for aerobic metabolism (fig 1). Oxygen consumption or uptake (Vo2) is determined by cellular requirements and is classically considered to be independent of ho2 unless the latter falls below a critical level (clbo2; fig 2A, point B). Below this point further reductions in DO2 result in a fall in Vo2.' The slope of the line relating uptake and delivery defines the oxygen extraction ratio (ERo2), which is an index of the efficiency of total tissue extraction of oxygen from the extracellular environment. Thus above cI)o2 Vo2 is supply independent and ERo2 falls progressively as DO2 rises. Below clto2 a state of so called supply dependency of oxygen uptake exists, the extent of which is determined by the slope of the line (ERo2). Individual organs have different values of ERo2, which may vary with stress and in different disease states. A more complete understanding of the global and regional 1)o2-Vo2 relationships would be a major advance in the management of the critically ill.5 TISSUE OXYGEN DELIVERY Oxygen delivery to individual cells is usually considered in terms of convective and diffusive oxygen transport. Convective oxygen transport refers to the bulk movement of oxygen in blood and includes the regional distribution of car- diac output to individual organs and the mechanisms regulating the tissue microcircu- lation. This is determined by a complex inter- action of endothelial, neural, and metabolic influences on arteries, small resistance arterioles, and precapillary sphincters.6 Never- theless, in the normal state and in clinical conditions convective transport depends predominantly on cardiac output. Diffusive oxygen transport refers to the transfer of oxygen molecules from blood through the extracellular matrix down the capillary-intracellular oxygen tension (Po2) gradient and it depends on arterial oxygen tension (Pao2). Analysis of DO2 Department of Intensive Care, St Thomas's Hospital, London SE1 7EH R M Leach D F Treacher Reprint requests to: Dr D F Treacher 971 on November 20, 2020 by guest. Protected by copyright. http://thorax.bmj.com/ Thorax: first published as 10.1136/thx.47.11.971 on 1 November 1992. Downloaded from
Transcript
Thorax 1992;47:971-978
Thepulmonaryphysician and critical care * 6 Series editor: T WEvans
Oxygen transport: the relation between oxygen
delivery and consumption
R M Leach, D F Treacher
The unifying concept of oxygen transport hasuntil recently been neglected by both car-diologists and respiratory physicians. With theincreasing part played by both disciplines inthe care of the critically ill, however, thisattitude is changing. The primary function ofthe heart and lungs is to generate a flow ofoxygenated blood to the tissues to sustainaerobic metabolism. The main requirements ofthis system are that it should be energy efficient,so that unnecessary cardiorespiratory work isavoided, and that it should be sensitive to thefluctuating demands of cellullar metabolism.Secondly, metabolic demand and distributionshould be matched regionally at rest, duringexercise, and in different disease states.Thirdly, oxygen should be able to pass efficien-tly across the extravascular tissue matrix. Themechanisms controlling oxygen distributionare incompletely understood, but are almostcertainly important in determining clinical out-come in the critically ill patient.'Although the relation between oxygen
delivery (Do2) and consumption (Vo2) has notbeen clearly established, these variables areoften measured to define a population ofcritically ill patients in whom Vo2 is limited byDo2, the state of so called "pathological supplydependency."2 During recent years many ofthe publications on critical care, and indeedpractice in leading intensive care units, haveemphasised the importance of raising Do2 to"supranormal" levels in an attempt to satisfythe increased metabolic demands of thesepatients. This practice has been justified by theobservation that increased DO2 improvesoxygen debt and outcome in postoperativesurgical patients requiring intensive care.3 Noone would dispute that restoring blood volumeto improve DO2 in the severely hypovolaemicpatient must be beneficial. Controlled trials,however, examining the influence of suchstrategies on clinical outcome in patients withmore complex conditions, suffering from sepsis,cardiovascular collapse, and hypoxic hypox-aemia, have produced conflicting data. Perhapsthe concept of global oxygen delivery has failedto emphasise the importance of the regionaldistribution of blood flow, particularly to thesplanchnic and renal vascular beds, which maybe more important in determining clinicaloutcome.'
This article reviews current ideas about therelation between DO2 and Vo2, the physio-
logical mechanisms controlling regional lo2,and their relevance to critical illness andcurrently used therapeutic interventions.
Physiology of oxygen transportGLOBAL OXYGEN DELIVERYUnder normal resting conditions global DO2(defined as the product of cardiac output andarterial oxygen content) is more than adequateto meet tissue oxygen demands for aerobicmetabolism (fig 1). Oxygen consumption oruptake (Vo2) is determined by cellularrequirements and is classically considered to beindependent of ho2 unless the latter falls belowa critical level (clbo2; fig 2A, point B). Below thispoint further reductions in DO2 result in a fall inVo2.' The slope of the line relating uptake anddelivery defines the oxygen extraction ratio(ERo2), which is an index of the efficiency oftotal tissue extraction of oxygen from theextracellular environment. Thus above cI)o2Vo2 is supply independent and ERo2 fallsprogressively as DO2 rises. Below clto2 a stateof so called supply dependency of oxygenuptake exists, the extent ofwhich is determinedby the slope of the line (ERo2). Individualorgans have different values of ERo2, whichmay vary with stress and in different diseasestates. A more complete understanding of theglobal and regional 1)o2-Vo2 relationshipswould be a major advance in the managementof the critically ill.5
TISSUE OXYGEN DELIVERYOxygen delivery to individual cells is usuallyconsidered in terms of convective and diffusiveoxygen transport. Convective oxygen transportrefers to the bulk movement ofoxygen in bloodand includes the regional distribution of car-diac output to individual organs and themechanisms regulating the tissue microcircu-lation. This is determined by a complex inter-action of endothelial, neural, and metabolicinfluences on arteries, small resistancearterioles, and precapillary sphincters.6 Never-theless, in the normal state and in clinicalconditions convective transport dependspredominantly on cardiac output. Diffusiveoxygen transport refers to the transfer ofoxygenmolecules from blood through the extracellularmatrix down the capillary-intracellular oxygentension (Po2) gradient and it depends onarterial oxygen tension (Pao2). Analysis of DO2
Department ofIntensive Care,St Thomas's Hospital,London SE1 7EHR M LeachD F TreacherReprint requests to:Dr D F Treacher
971
on Novem
ber 20, 2020 by guest. Protected by copyright.
http://thorax.bmj.com
/T
horax: first published as 10.1136/thx.47.11.971 on 1 Novem
should evaluate both mechanisms as failure ofeither component results in tissue hypoxia.The relative importance of convective and
diffusive oxygen transport in limiting Vo2 iscontroversial. As the oxygen content of bloodperfusing a limb at different flow rates isgradually reduced, Vo2 and ERo2 are greaterfor a given Do2 under low flow conditions.7 Thissuggests that Vo2 is diffusion limited rather thanconvection limited in progressive hypoxaemia.In contrast, other workers have shown that thepoint at which Vo2 falls is similar whether Do2is decreased by progressive anaemia, hypox-aemia, or a reduction in blood flow, and that Vo2
Figure 2 A- The normal(classical) oxygen delivery(D2)-oxygen uptake(Vo2) relationship. PointB represents the criticalDo2, A-B the supplydependent phase, B-C thesupply independent phase.B-"Pathological" supplydependency; A-B' shows acurvillinear relationship.At high levels ofDo2 doesthe relationship remainsupply dependent (B'-D)or become independent(B'-C')?
200
100
300
200
100
A
NORMAL(classical)
B C
A
correlates with Do2 but not with venous Po2 .48These observations would suggest a dominantrole for convective transport. To clarify thisparadox the theoretical relation between tissueoxygen supply and uptake was examined re-cently by using the Krogh tissue cylindermodel9 (fig 3). In this hypothetical model,provided that the intercapillary distances wereless than 80 ,sm, Vo2 began to fall at a similarlevel of Do2 (cDo2), independent of whetherDo2 was progressively reduced by anaemia,hypoxaemia, or a fall in blood flow. As theintercapillary distances were progressivelyincreased above 80 ,um there was a rightwardshift in c1o2 and a flattening of the responsegradient. These changes in the Vo2-Do2relationship were more dramatic when thereduction in Do2 was achieved by hypoxaemiathan with anaemia or a fall in flow. Thissuggests that the relative contributions of con-vective and diffusive transport are determinedby capillary density, which may vary betweenorgans, with tissue oedema, and with diseasestates. This theoretical analysis may explain theearlier contradictory results.
400 800 1200I
, _.-IDo2 (ml min - m-l) Relation between oxygen supply andoxygen consumptionDefining the Do2-Vo2 relationship in individualsubjects over a sufficiently wide range of values
CRITICALLY ILL is extremely difficult for practical and ethicalB D reasons. Consequently, plotting complete Do2-
D Vo2 relationships has been impossible, par-BP ?__ ticularly in healthy individuals, though clinical
_/ C' requirements have permitted some progress toB/ ....... C be made in patients with certain disease states.
This lack of data has generated many hypo-theses, but no scientifically secure conclusionregarding the phenomenon of supply depen-dency of oxygen uptake. Our present under-
[A / standing is founded on evidence from (a)400 800 1200 exercise in healthy human volunteers, (b)
-1-2) animal experiments and patients undergoingDo2 (ml mn- m ) surgery, and (c) critically ill patients.
}RQ
Ic
E
N0
E
0
972
on Novem
ber 20, 2020 by guest. Protected by copyright.
http://thorax.bmj.com
/T
horax: first published as 10.1136/thx.47.11.971 on 1 Novem
Oxygen transport: the relation between oxygen delivery and consumption
Figure 3 Predicted effectof changes inintercapillary distance(perfused capillarydensity) on critical oxygendelivery. Reductions inoxygen delivery wereachieved by (A)progressive anaemia, (B)hypoxia, (C) stagnanthypoxia. The numbers onthe plots correspond to theintercapillary distances inmicrometres. Whenintercapillary distanceswere short (40-80 jmj,similar levels of criticaloxygen delivery were seenfor all three forms ofhypoxia. Whenintercapillary distanceswere over 80 jm tissueoxygen uptake (VO) wasreduced to a greater extentby hypoxic hypoxia thanby anaemic or stagnanthypoxia. (Reproducedfrom Schumacker andSamsel by courtesy of theJournal of AppliedPhysiology.9)
7'
cm
Ilec
EE
0
c
le
a)x0
Ele
E6aI2
CLcs91)a)
x0
6
5
4
3
2
0
7
6
5
4
3
2
0
7
a)
c
0.
a)
2-
0x0
6
5
4
3
2
A
80 160
/
240
An-emic the anaerobic threshold is about 0 6, whichAnaemic implies that above this level oxygen uptake
becomes diffusion limited." ERo2 is improvedby training, which suggests an improvement inthe diffusion reserve (resulting from increasednumbers of mitochondria or microvessels) ormore efficient regional distribution ofblood flowto the active tissues.'2
EVIDENCE FROM ANIMALS AND PATIENTSUNDERGOING SURGERY
5 I0 I5 | The response of the cardiopulmonary system5 10 15 20 to decreasing oxygen availability is more com-
plex. The evidence for the classical two phaserelationship between Vo2 and DO2 (fig 2A)4 restsalmost exclusively on animal studies. Thesehave shown that cDo2 varies from 6 to 10 ml/
B Hypoxic kg/min, with an ERo2 of0 5-0 8.'3 '4 In the only40 80 study in relatively healthy human subjects
under normal conditions, a biphasic relation-ship between Vo2 and DO2 was observed in 58patients undergoing coronary artery bypass160
/ 200-/ surgery after the induction of anaesthesia, but200
// before the start of cardiac bypass.'5 Analysis of240 99 DO22/Vo2 points revealed that Vo2 was con-///s stant for values of Do2 from 330 to 700 ml
mi' m 2 Below the cD98.433O...r..mL-min'L____________=_______________________ m-2), which was similar to that found in animal
5 10 15 20 studies, Vo2 fell with decrea The V02in these patients was substantially lower thanthe basal levels recorded in animals and thiswas attributed to anaesthesia. At the cDo2 theapparent Eo2-was only-033, a figure well
r+5nont below that observed in animal experiments.wtagnant40.
/~240
0 5 10 15 20
Oxygen delivery (ml mmi n - kg
EVIDENCE FROM EXERCISE IN HEALTHY HUMANVOLUNTEERS
The cardiorespiratory system is extremelyeffective at augmenting DO2 to satisfy suddenincreases in metabolic demand required duringexercise. Immediate increases in cardiac outputcan rapidly increase the basal DO2 five fold inthe healthy individual. As Vo2 may increase 10
fold, however, during maximal exercise andERo2 also rises at the onset of exercise, a
progressive widening of the arteriovenousoxygen difference occurs."'' At rest the ERo2 is
about 0 3 but can increase to 0 8 during maxi-mal exercise in healthy, well trained subjects.This is attributed to an improvement in dif-fusive oxygen transport mediated by increasesin nutritive capillary densities. When themetabolic demands of increasing exercise canno longer be sustained by increases in Do2 or
ERo2 the anaerobic threshold is reached andblood lactate concentrations rise. The ERo2 at
EVIDENCE FROM CRITICALLY ILL PATIENTSIn the critically ill most investigators havereported a modified Do2-Vo2 relationship, sug-gesting thait under pathological conditions DO2and Vo2 continue to increase above thephysiological level for cDo2 (fig 2B). ThereforeVo2yremapply limited at a normal or even'increased Do2. In addition, ERo2 does notincrease, as is observed during exericse, or withprogresive reductions in Do2, as reported inanima relting singlephase, curvilinear relationship is termedpathological supply dependency (fig 2B) andiffiplies that a reduction in Do cannot becompensated for by an increase in ERo Vtherefore must fall. This relationship has beenreported both in animal models of sepsis andduring critical illness in man. It was firstobserved in patients with the adult respiratorydistress syndrome (ARDS)'6"' and was sub-sequently identified in patients with septic andhypovolaemic shock,'8 19 congestive cardi'cfailure,2C pulmonary hypertension,2' andchronic obstructive airways disease.22 Thelinear relationship has important clinicalimplications, because with a ritaifand downward displacement of the Vo2-Dorelationsip, as illustrated in figure 2B, Vo2 isreduced at a normal Do. At a simplistic level,rthis suggests that an increase in total DO2 bymanipulation of the circulation may be bene-ficial. If, however, Vo2 is diffusion limitedtreatment should be directed at improving Pao2rather than increasing DO2. Factors affectingthe gradient and movement of the Do2-Vo2
u i a I
973
I
1_
,\
on Novem
ber 20, 2020 by guest. Protected by copyright.
http://thorax.bmj.com
/T
horax: first published as 10.1136/thx.47.11.971 on 1 Novem
Figure 4 Hypotheticaloxygen delivery (Do2)-oxygen uptakel ( V02)relationship. Point Arepresents the averagevaluesfor Do2 Vo2foundin normal subjects at rest.'0To the right ofpoint A liesthe regression linefitted tothe values of DQ6 and Vn2from the progressiveexercise data of Astrandet al'°; it represents thecondition where Vo2 is theindependent variable asDo2 rses to meet increasedcellular energyrequirements and shows theincreasing oxygenextraction ratio (ERo2).To the left of point A is theportion of the relationshipwhere Vo2 is the dependentvariable, which is moredifficult to define because ofthe lack of datacomparable to that onexercise in healthy humansubjects. The actual ERo2in normal man is unknownand three possiblepathways for differentERo2 values (0 33, 0 50,and 0 70) are illustrated.(ReproducedfromDantzker et al by courtesyof the American Review ofRespiratory Disease.23)
200 400 600 800Oxygen transport (ml min-' m- 2)
relationship in health and disease clearlyrequire further investigation to determine theprognostic significance and implications ofsuchchanges. The basis for the belief that thecurvilinear relationship identified in patho-logical supp dependency is distinctfrom theclassical biphasic relationship reported inhealth is tenuous.
IS THE Vo2-6o2 RELATIONSHIP BIPHASIC ORLINEAR?
Despite the biphasic Vo2-Do2 relationship inanimal experiments, it has recently been sug-gested that in man both the physiological andthe pathological relationships are linear.23 In arecent review Dantzker and colleagues pro-posed the hypothetical interaction of DO2 andVo2 shown in figure 4. Experiments duringexercise provide strong evidence to supportthat part of the hypothetical line to the right ofpoint A.'° ATi-
shipiwould demand ERo2 values of 0 5-07 toachieve the plateau phase. In human studies,how mum values for ERo2 both indisease states (0 25-045) and in relative health(0-33)'5 are much lower. These low ERo2 valuesargue against the existence of the point cDo2
and a plateau phase. If a- curviltinear relation-ship-s-imilarTo That observed in "pathologicalsupply dependency" existed in healthy indi-viduals this would satisfy the requirement forcardiorespiratory efficiency as it implies a muchsmaller reserve of oxygen transport.23An alternative explanation is that the
curvilinear relationship between supply anddemand represents the distillation of a series ofpoints lying on a family of organ specificsupply-demand curves. hese points would bedetermined by the regional distribution of flowbetween orga hing diffcrcnt and varying Vo2-Do2 relationships.
Clinical implications of the globalD02-Vo2 relationshipPatients admitted to an intensive care unit whoshow the pattern of pathological supply depen-dency have a 70% mortality, compared with a30% mo a milarly,
in mechanically ventilate'd'patients with sepsis
the prognosis was found to be worse if Vo2increased in response to a rise inDo2 producedby an infusion o prosracyclin (that is, patho-~logical supply dependency).25 Shoemaker et al,3in a large series ofpostoperative patients admit-
ted to intensive care, empirically set "supra-normal" Do2 and Vo, targ and cported-animproveientirirruutcome.This evidence has resulted in the current
vogue for "goal directed" treatment requiringthat the cardiac index should be >4 5 1min- m2, Do2 > 600 ml min' m-2, Vo2 >170ml min-m' , and venous oxyen s,tuation(Svo2) >70%.339 is approach to manage-ment assumes that the following conditionspertain: (a) Vo2 can be increased by raising Do2to supranormal levels; (b) an increase in VO2reduces tissue hypoxia; (c) an increase in Do2improves outcome. The validity of thesestatements depends on the individual patho-logical process and whether there is a truedeficit of either convective or diffusive oxygentransport. The importance of these two factorsin oxygen delivery is examined by consideringtwo groups of patients.
STUDIES IN PATIENTS WITH CHRONICOBSTRUCTIVE AIRWAYS DISEASEPatients with hypoxic chronic obstructive air-ways disease have b6tT a-fail`re 6F-ar-teriHaloxygenation and abnormalities of cardiac out-put. Pathological supply dependency has beenreported in both chronic obstructive airwaysdisease22 and pulmonary hypertension.2' Threeyears after the development of peripheraloedema in hypoxaer-i'c-or-pumuonale survivalis less than 50%. Ca-n this reduced survival beattribu,teifto compromisd oxvgen diffusiondue to reduced Pao2 or to reduced DO2 from afailure of convective transport? In a studyexamining 50 patients with severe chronicobstructive airways disease DO2 and the cardiacindex had no prognostic significance, whereas aclear benefit was associated with increased Pao,and venoUs oxygen tesion-(Pvo2).27 Acuteoxygen therapy produces a fall in cardiac indexand Do2 but an increase in Pvo2, implyingimproved tissue oxygenation.27 Long termoxygen therapy aimed at improving arterialoxygen s atces no sig-nificant haemodynamic benefit or increase inDO2, but does improve survival in patients withypoxaemic cor pulmonale. Reversal of arterialhypoxaemia (>.71 kPa) appears to be the keyprognostic factor.262829 This suggests that forthis population of patients the diffusive com-ponent of oxygen transport is more importantthan the convective component.
STUDIES IN CRITICALLY ILL PATIENTSThe position in the critically ill appears to bedifferent. Several reports have suggested thatimproving o,ydincreasing Do2 to a point wellabove c1)o2 may confer clinical beneht. 3Empirical studie andcolleagues showed that increasing cardiac out-put, Do2, Vo2, and blood volume above normalin postoperative patients improved outcome.3Maintaining a high Do2 also improves Vo2 andreverses lactic acid-production.32 I nese studiessugges't that a high DO2 should be maintainedwhenever possible, but this generalisation can-not allow for either specific aspects of theunderlying pathology or the possible complica-tions of strategies used to increase DO2. In
974
on Novem
ber 20, 2020 by guest. Protected by copyright.
http://thorax.bmj.com
/T
horax: first published as 10.1136/thx.47.11.971 on 1 Novem
Oxygen transport: the relation between oxygen delivery and consumption
ARDS attempts to increase Do2 by raisingfilling pressures with volume expansion willhave detrimental effects due to increasedalveolar capillary permeability. Any increase inpulmonary artery occlusion pressure willincrease lung water, further impairing gasexchange and lung compliance and leading toan increase in the fractional inspired oxygen(Fio2) requirement and airway pressures. Thesame increase in vascular permeability, albeitless clinically apparent, affects all tissues.Therefore volume loading may increase extra-vascular tissue water and potentially com-promise diffusive oxygen transport.33A second consideration in ARDS is the
frequent need for positive end expiratory pres-sure (PEEP) to reduce the alveolar-arterialoxygen gradient and FIO2 requirements. Theeffect of increasing PE P on DO2 is unpredict-able but usually a net fall occurs owing toreduced cardiac output even if Pao2 rises. Thispattern of response provides an interestingmodel with which to study the relative con-tribution of diffusive and convective transport.Carlile and Gray investigated the effect ofopposite changes in cardiac output and Pao2 onDo2, Vo2 and Pvo2 and showed that in thesecircumstances Vo2 was not supply dependent.34
MEASUREMENT ERRORS: MATHEMATICAL ANDPHYSIOLOGICAL LINKAGEThe conflicting evidence concerning supplydependency in the critically ill may be ex-plained by the different methods used formeasuring Vo2. Studies in patients after sur-gery and in those with sepsis and ARDS haveshown that, although DO2 correlated well withVo2 calculated from cardiac output and dif-ference in arterioveno'-s oxygenucLneCnt (fin-direct Fick method), the correIation was muchweaker when Vo2 was calculated from directmeasurements of ventilatory minute volumeand inspired and mixed expired oxygen con-centration by mass spectrometer or metaboliccart.3937 It was suggested that the Do2/Vo2relationship seen when Vo2 is calculated by theindirect Fick method is an artefact related tothe use of common variables (cardiac outputand arterial oxygen content) in the calculationof both Vo2 and DO2 .38 Errors in measuring thetwo variables are therefore effectively coupled,producing apparent supply dependency. Theeffect ofthis mathematical coupling depends onthe size of the errors in the measurement of thevariables concerned.39 In studies with a smallrange of DO2 the problems are exaggerated.39 Ifmeasurement errors are large, as may be thecase during mechanical ventilation, the chancethat a true athological supply dependencyexists is di d ic ehowever, suggests that the lnar relioipbetween Do2 and Vo2 cannot be ascridenti t atemati
Physiological coupling may also be responsi-ble for generating a false association betweenvariables. If, for example, inotropic support isused to increase Do2, most of the agentsavailable not only increase cardiac output butalso stimulate metabolism, thereby increasingVo2 directly.' Data obtai needin such circum-stan--es-muld be misinteroreted as evidence ofsupply dependency. uch treatment could
hardly be viewed as beneficial as it is similar tofever in its effect on Vo2/Do2. Pain, anxiety,endogenous catecholamine release, andphysical activity produce similar changes.Doubts about a simple global supply depen-
dency relationship have also been raised bypersonal observations in patients with severesepsis characterised by systemic hypotensionwith high cardiac output and Do2. In thesepatients with much reduced systemic vascularresistance treatment with noradrenaline in-creased not only arterial pressure but alsn'heincidence of clinicalndices of organ function.Independent measurements ot V02 by indirectcalorimetry using a mass spectrometer and oTDO2 from tnermodilurion cardiac output andarterial oxygen content showed that, with theintroduction of noradrenaline,rwhileDo2-mained unchauiged- u vTis fell, Vo2 increased(DFT, unpublished observations). This evi-dence oT an inverse relation between bO2 andVo2 in certain circumstances could be ex-plained by regional redistribution of flow andwould suggest that such changes may be moreimportant than absolute increases in Do2.
Clinical implications of regional flo2-Vo2relationshipsIn the critically ill patient the regional distribu-tion to the organs of the total oxygen deliveredvaries considerably with the underlying patho-logical process. Both clinical and theoreticalevidence shows that Pao2 is more importantthan DO2 in malntaining tissue oxygenation,particrly in respiratory-fa-i with hypox-aemia.79 In these cir-umstances treatment thatimproves peripheral distribution and cellularutilisationot oxygen is more appropriate.
Induction of endotoxaemia or septicaemia inanimal experiments has shown that the cf)o2 ishigher and the ERo2 reduced to a greater extentin-the splanchnic than instneskeletal musclecirculation. Splanchnic perfusion is selectivelyredfied by the endogenous vasoconstrictorsreleased in critical illness and the gut mucosa isfurther compromised by the frequent failure tosustain enterafteeding. The fall il nwith increasde metabolic demands and reducedERo2 results in ischaemia of the splanchnicorgans, particularly the gut. This renders thegut mucosa "leaky," allowing absorption ofendotoxin and translocation of bacteria into theportal cir smeans that a massive toxic load enters theportal circulation, which first overwhelms theliver defences and t d'prmoires illaryendothelial damage in the lung (ARDS) and thedevelopment ofmultisystem organ failure. Themost common cause of late death in the inten-sive care unit ismu lu andrecent evidence has shown that treatmentaimed at maintaining or improving spTanchnicperfusion and tissue oxygenation reduces mor-tality and improves outcome.'
Therapeutic strategies to manipulateD02-Vo2 relationshipsINOTROPIC SUPPORTCatecholamines, including dobutamine andadrenaline, improve cardiac output and are
975
on Novem
ber 20, 2020 by guest. Protected by copyright.
http://thorax.bmj.com
/T
horax: first published as 10.1136/thx.47.11.971 on 1 Novem
reported to improve survival in septic shock.4'In addition, they may have important effects onregional vascular resistance, thereby improvingoxygen diffusion and tissue ERo2 . Dobutamine,however, may reduce splanchnic perfusion atthe doses required to increase global DO insepsis, emphasising the desirability of measur-inig regional perfusion iftherapeutic benefits areto be assured.' Inotropes may also inducephysiological linkage, increasing cardiac workan and d n-ergiCagent dopexamine hydrochloride has no oaadrenergic effects and may selectively increaserenal and splanchnic blood flow.44 45 The poten-tiaFvalue of region specific inotropic agents isconsiderable but needs further evaluation.
VOLUME INFUSIONMany trials have examined the effects ofvolumeloading to increase cardiac output by raisingfilling pressures. Volume loading has beenshown to increase cardiac output and total Do2and to improve disease outcome post-operatively.3 Nevertheless, there are two majorproblems with the indiscriminate use of thisstrategy. Firstly, complications occur inpatients with increased vascular permeability,as already discussed. Secondly, many criticallyill patients? particularly those with sepsis, haveseverely impaired myocardial contractility,with flat myocardial tunction curves, so that thestroke volume and hence gain in cardiac outputfrom an increase in filling pressure is minimal.This results in all the disadvantages of volumeloading in exchange for trivial improvements inDo2 .
VASODILATORSVasodilators increase oxygen transport withoutincreasing cardiac work or producing theproblems of physiological linkage,46 but inmany critically ill patients systemic vascularresistance is already low, and poor organ per-fusion resulting from systemic hypotension is amajor concern. Even if vasodilators are toler-ated without overt haemodynamic compro-mise, their effects on regional distribution areunpredictable and in certain circumstancesmay jeopardise blood flow in vital organs, eventhough overall DO2 is increased.46 The use ofvasodilators to re ea covert oxygen debt byshowing whether Vo, increases in responsetoDO2 has been recommended,2nbut no convinc-ing evidence exists to show that vasodilatorsimprove outcome.
PERIPHERAL CIRCULATIONManoeuvres that shift the oxygen dissociationcurve to the right will increase tissue oxygenuptake. In this regard the adverse effects of araised temperature and acidosis are well knownbut the importance of correcting hypophos-phataemia is frequently overlooked.47Most methods for improving tissue oxygena-
tion by influencing the microcirculation are atpresent experimental. The endothelium hasbeen shown to have an important influence onvascular tone and vasoreactivity through therelease of both constricting and relaxing fac-tors. The endothelial cell damage associated
with septic shock andARDS is likely to have animportant influence on tissue perfusion. Themost important of the endothelium derivedrelaxing factors is nitric oxide. Recent studieshave shown that inhibition of the synthesis ofendothelium derived relaxing factors may bebeneficial in maintaining vascular tone in septicshock.48 Anti titlybeen shown to reduce mortality in patients withGram particuarlythose with septic shock, though the mechanismof action in terms of oxygen transport andregional perfusion is unclear.4 Other anti-oxidants, such as prostaglandin E,, mayimprove 1)02 and Vo2 but have no influence onmortality.50
FACTORS INFLUENCING V02It is important to avoid factors which raisemetabolic demand, with an inevitable increasein Vo2 and DO2, but which may aggravaterather than relieve tissue hypoxia. Principalamong these are infection, a high temperature,drugs (f agonists), excess physical activity(physiotherapy, restlessness, shivering, fightingventilator), sympaThetic overactivity (pain,anxiety), and the feeding regimen (excess intra-venous glucose).
Measures of the adequacy of tissueoxygenationLACTATE CONCENTRATIONBlood lactate concentration may be raised ornormal in the presence or absence of hypoxiabecause the metabolic pathways utilisingglucose during aerobic energy production maybe blocked at several points.5' If phosphofruc-tokinase is inhibited, glucoseoUisz6riTisprevented without an increase of lactate. Incontrast, in the hypermetablism of sepsisinactivation of pyruvate dehydrogenaseprevents utilisation of pyruvate in the citricacid cycle, resulting in production of lactate,accumulation of pyruvate, and aerobic meta-bolism of fat and protein.52 Recent studies haveshown that endotoxin can directly inactivatepyruvate dehydrogenase, resulting in lactateproduction in the absence of tissue hypoxia.3With an unfavourable cellular redox state,normal Do2 may be associated with high lactateconcentrations, but on the other hand if com-pensatory reductions in [ATP]/[ADP][Pi] or[NAD+]/[NADH] occur then low lactate con-centrations may be found with tissue hypoxia.54Lactate is therefore not a reliable reflection oftissue hypoxia.The blood lactate concentration also
represents a balance between production, asshown by perfusion, and consumption byhepatic, cardiac, and skeletal muscle metabol-ism.2 Clinically, arterial lactate concentrationsare reported to vary inversely with DO22,5 butthe suggestion that pathological supply depen-dency occurs only when blood lactate concen-trations are raised is incorrect as the samerelationship may be found in patients withnormal lactate concentrations.20 Clearly, thevalue of a single lactate concentration in theassessment of tissue hypoxia is at best question-
976
on Novem
ber 20, 2020 by guest. Protected by copyright.
http://thorax.bmj.com
/T
horax: first published as 10.1136/thx.47.11.971 on 1 Novem
Oxygen transport: the relation between oxygen delivery and consumption
able.55 Serial lactate measurements, par-ticularly if corrected for pyruvate, may be ofgreater value.
TONOMETRYMuch of the evidence presented here hassuggested the probable importance of regionaldistribution of Do2, particularly to the splan-chnic bed. The introduction of the gastro-intestinal tonometer to measure the intra-mucosal pH t (pHi) promises to be animportant advance. 'his may provide an earlywarning-of--irmdequate splanchnic tissueoxygenation and guide subsequent treatment.56Recent reports have shown that a low pHi isassociated with a poor prognosis, but if pHiincreases with treatment the outcome is im-proved.'
Future developmentsFurther advances in the understanding ofoxygen transport require the development ofsuitable techniques for measuring regionalblood flow and tissue oxygenation. Examinationof serial muscle biopsy specimens to study thehistological changes, endothelial architecture,lactate concentrations, and ATP turnover mayhelp to explain the microcirculatory changescontrolling tissue perfusion.57 Magnetic reson-ance spectroscopy is a non-invasive method bywhich intral bic and anaerobicenergy metabolism may be studied. Phos-phorus-3 1 is a naturally occurring isotopeconcerned with energy transfer within the cell.3'P-magnetic resonance spectroscopy generatesspectral peaks corresponding to the resonanceof the phosphate bons TP,inorganic phosphate (P.), and phosphocreatine(PCr), from which intracellular pH (pHi),oxidative phosphorylation, the rate of mito-chondrial electron transport, and creatine kin-ase kinetics may be calculated.5" Despiteappreciable technical difficulties and the prac-tical problems presented by patients in inten-sive care, magnetic resonance spectroscopypromises considerable advances in the study ofboth aerobic and anaerobic metabolism and themetabolic response to critical illness. Positronemission tomography permits the non-invasivemeasurement of regional blood flow and organfunction and should allow the response of themicrocirculation to stress, hypoxaemia, andtreatment to be assessed.60
ConclusionDespite a bewildering array of publications,many aspects of oxygen transport remain anenigma. The relationship between DO2 and Vo2has not been'ctndvogue of therapeutic manipulation is ofunproved value in determining outcomemany conditions. Part of the confusion relatesto problem;s in measuring Vo, and futurestudies should report independently calculatedvalues from the analysis of inspired and expiredgas. In critically ill patients it is essential todefine clearly the population and the path-ological abnormality under study and to iden-
tify and account for potentially confoundingmetabolic factors. The recognition that tissuehypoxia varies between and within individualorgans is focusing attention on the importanceof regional oxygen delivery and the need fordirect assessment of tissue oxygenation.6'Recent technological advances should allowfurther aspects of the enigma to be unravelled.
I Doglio GR, Pusajo JF, Egurrola MA, et al. Gastric mucosalpH as a prognostic index of mortality in critically illpatients. Crit Care Med 1991;19:1037-40.
2 Vincent J-L, Roman A, De Backer D, Kahn RJ. Oxygenuptake/supply dependency. Am Rev Respir Dis 1990;142:2-7.
3 Shoemaker WC, Appel PL, Kram HB, Waxman K, LeeT-S. Prospective trial of supranormal values of survivorsas therapeutic goals in high risk surgical patients. Chest1988;94:1 176-86.
4 Cain SM. Oxygen delivery and uptake in dogs during anemicand hypoxic hypoxia. J Appl Physiol 1977;42:228-34.
5 Nelson DP, Samsel RW, Wood LDH, Schumacker PT.Pathological supply dependence of systemic and intestinal0, uptake during endotoxaemia. J Appl Phvsiol 1988;64:2410-9.
7 Gutierrez G, Pohil RJ, Strong R. Effect of flow on 0°consumption during progressive hypoxaemia. J AppiPhvsiol 1988;65:601-7.
8 Schumacker PT, Long GR, Wood LDH. Tissue oxygenextraction during hypovolaemia: role of hemoglobin P50.J Appl Physiol 1987;62:1801-7.
9 Schumacker PT, Samsel RW. Analysis of oxygen deliveryand uptake relationships in the Krogh tissue model.J Appl Physiol 1989;67:1234-44.
10 Astrand P, Cuddy TE, Saltin B, Stenberg J. Cardiac outputduring submaximal and maximal work. J Appi Phvsiol1964;19:268-74.
11 Weber KT, Janicki JS. Cardiopulmonary testing for evalua-tion of chronic cardiac failure. Am J Cardiol 1985;55:22-31A.
12 Ekblom B, Astrand P, Saltin B, Stenberg J, Wallstrom B.Effect of training on circulatory response to exercise.J Appl Physiol 1968;24:518-28.
13 Gutierrez G, Warley AR, Dantzker DR. Oxygen deliveryand utilisation in hypothermic dogs. J Appl Phvsiol 1986;60:751-7.
14 Nelson DP, Beyer C, Samsel RW, Wood LDH, SchumackerPT. Pathological supply dependence of 0° uptake duringbacteraemia in dogs. J Appl Physiol 1987;63:487-92.
15 Shibutani K, Komatsu T, Kubal K, Sanchalla V, Kumar V,Bizzarri DV. Critical level of oxygen delivery in anaes-thetized man. Crit Care Med 1983;li:640-3.
16 Powers SR, Mannal R, Neclerio M, English M, Marr G,Leather R, et al. Physiological consequences of positiveend-expiratory pressure (PEEP) ventilation. Ann Surg1973;178:265-72.
17 Danek SJ, Lynch JP, Weg JG, Dantzker DR. The depen-dence of oxygen uptake on oxygen delivery in the adultrespiratory distress syndrome. Am Rev Respir Dis 1980;122:387-95.
18 Kaufman BS, Rackow EC, Falk JL. The relationshipbetween oxygen delivery and consumption during fluidresuscitation ofhypovolemic and septic shock. Chest 1984;85:336-40.
20 Mohsenifar Z, Amin D, Jasper AC, Shah PK, Koerner SK.Dependence of oxygen consumption on oxygen delivery inpatients with chronic congestive heart failure. Chest1987;92:447-56.
21 Mohsenifar Z, Jasper AC, Koerner SK. Relationship be-tween oxygen uptake and oxygen delivery in patients withpulmonary hypertension. Am Rev Respir Dis 1988;138:69-73.
22 Brent BN, Matthay R, Mahler DA, Berger HJ, Zaret BL,Lister G. Relationship between oxygen uptake and oxygentransport in stable patients with chronic obstructivepulmonary disease. Am Rev, Respir Dis 1984;129:682-6.
23 Dantzker DR, Foresman B, Gutierrez G. Oxygen supplyand utilisation relationships. A reevaluation. Am RevRespir Dis 199 1;143:675-9.
24 Gutierrez G, Pohil RJ. Oxygen consumption is linearl,related to 0. supply in criticallv ill patients. J Crit Care1986;1:45-53.
25 Bihari D, Smithies M, Gimson A, Tinker J. The effect ofvasodilation with prostacyclin on oxygen delivery anduptake in critically ill patients. N Enigl J Med 1987;317:397-403.
977
on Novem
ber 20, 2020 by guest. Protected by copyright.
http://thorax.bmj.com
/T
horax: first published as 10.1136/thx.47.11.971 on 1 Novem
26 Medical Research Council. Long term domiciliary oxygentherapy in chronic hypoxaemic cor pulmonale complicat-ing bronchitis and emphysema. Report ofa working party.Lancet 1981;i:681-6.
27 Kawakami Y, Kishi F, Yamamoto H, Mtyamoto K. Relationof oxygen delivery, mixed venous oxygenation, and pul-monary hemodynamics to prognosis in chronic obstruc-tive pulmonary disease. N Engl J Med 1983;308:1045-9.
28 Nocturnal Oxygen Therapy Trial Group. Continuous ornocturnal oxygen therapy in hypoxaemic chronic obstruc-tive lung disease: a clinical trial. Ann Intern Med 1980;93:391-8.
29 Cooper CB, Waterhouse J, Howard P.Twelve year clinicalstudy of patients with hypoxic cor pulmonale given longterm domiciliary oxygen therapy. Thorax 1987;42:105-10.
30 Lorente JA, Renes E, Gomez-Aguinaga MA, Landin L,DeLa Morena JL, Liste D. Oxygen delivery dependentoxygen consumption in acute respiratory failure. Crit CareMed 1991;19:770-5.
31 Edwards JD. Oxygen transport in cardiogenic and septicshock. Crit Care Med 1991;19:658-64.
32 Edwards JD. Practical application of oxygen transportprinciples. Crit Care Med 1990;18:545-8.
33 D'Orio V, Mendes P, Carlier P, Fatemi M, Marcelle R.Lung fluid dynamics and supply dependency of oxygenuptake during experimental endotoxic shock and volumeresuscitation. Crit Care Med 1991;19:955-62.
34 Carlile PV, Gray BA. Effect of opposite changes in cardiacoutput and arterial PO2 on the relationship between mixedvenous P02 and oxygen transport. Am Rev Respir Dis1989;140:891-6.
35 Vermeij CG, Feenstra BWA, Bruining HA. Oxygen deliveryand oxygen uptake in postoperative and septic patients.Chest 1990;98:415-20.
36 Ronco JJ, Phang PT, Walley KR, Wiggs B, Fenwick JC,Russel JA. Oxygen consurnption is independent ofchanges in oxygen delivery in severe adult respiratorydistress syndrome. Am Rev Respir Dis 1991;143:1267-73.
37 Bizouarn P, Soulard D, Blanloeil Y, Guillet A, Goarin Y.Oxygen consumption after cardiac surgery-a comparisonbetween calculation by Fick's principle and measurementby indirect calorimetry. Intens Care Med 1992;18:206-9.
38 Archie JP. Mathematical coupling: a common source oferror. Ann Surg 1981;193:296-303.
39 Statton HH, Feustel PJ, Newell JC. Regression ofcalculatedvariables in the presence of shared measurement error.J Appi Physiol 1987;62:2083-93.
40 Green CJ, Fraser RS, Underhill S, Maycock P, FairhurstJA, Campbell IT. Metabolic effects of dobutamine innormal man. Clin Sci 1992;82:77-83.
41 Edwards JD, Brown GCS, Nightingale P, Slater RM,Faragher EB. Use of survivors' cardiorespiratory values astherapeutic goals in septic shock. Crit Care Med 1989;17:1098-103.
42 Fellows IW, Bennett T, MacDonald IA. The effects ofadrenaline upon cardiovascular and metabolic functions inman. Clin SCi 1985;69:215-22.
43 Bouffard Y, Tissot S, Viale JP, et al. The effects of norepine-phrine infusion on oxygen consumption in a patient with
septic shock. Intens Care Med 1990;16:133-4.44 Lokhandwala MF, Jandhyala BS. Effects of dopaminergic
agonists on organ blood flow and function. Clin Intens Care1992;3(suppl): 12-5.
45 Stephan H, Sonnatag H, Henning H, Yoshimine K.Cardiovascular and renal haemodynamic effects of dopex-amine: comparison with dopamine. Brit J Anaesth 1990;65:380-7.
46 Leier CV. Regional blood flow response to vasodilators andinotropes in congestive heart failure. Am J Cardiol 1988;62:86-93E.
47 Clerbaux T, Detry B, Reynaert M, Kreuzer F, Frans A.Reestimation of the effects of inorganic phosphates on theequilibrium between oxygen and hemoglobin. Intens CareMed 1992;18:222-5.
48 Petros A, Bennet A, Vallance P. Effect ofnitric oxide synthaseinhibitors on hypotension in patients with septic shock.Lancet 1991;338:1557-8.
49 Ziegler EJ, Fisher CJ, Sprung CL, et al. Treatment ofgram-negative bacteremia and septic shock with HA-IA humanmonoclonal antibody against endotoxin: a randomised,double blind, placebo-controlled trial. N Engl J Med1991 ;324:429-36.
50 Bone RC, Slotman G, Maunder R, et al. Randomised doubleblind multicentre study of prostaglandin E, in patientswith the adult respiratory distress syndrome. Chest 1989;96:114-9.
51 Cain SM, Curtis SE. Experimental models of pathologicoxygen supply dependence. Crit Care Med 1991 ;19:603-11.
53 Vary TC, Siegel JH, Nakatani T, et al. Effects of sepsis onactivity of pyruvate dehydrogenase complex in skeletalmuscle and liver. Am J Physiol 1986;250:E634-9.
54 Wilson DF, Erecinska M, Drown C, Silver IA. Effect ofoxygen tension on cellular energetics. Am J Physiol 1977;233:C135-40.
55 Silverman HJ. Lack of a relationship between inducedchanges in oxygen consumption and changes in lactatelevels. Chest 1991;100:1012-5.
56 Landow L, Phillips DA, Heard SO, Prevost D, VandersalmTJ, Fink MP. Gastric tonometry and venous oximetry incardiac surgery patients. Crit Care Med 1991;19:1226-33.
57 Okada K. Reversible changes of skeletal muscular capillariesafter application of a tourniquet. Microvasc Res 1990;39:156-68.
58 Gutierrez G, Palizas F, Marini CE. Cellular energy metabol-ism. Recent advances in the study of the diaphragm withmagnetic resonance spectroscopy. Chest 1990;97:975-82.
60 Schuster DP. Positron emission tomography: theory and itsapplication to the study of lung disease. Am Rev Respir Dis1989;139:818-40.
61 Harrison DK, Kessler M, Knauf SK. Regulation ofcapillaryblood flow and oxygen supply in skeletal muscle in dogsduring hypoxaemia. J Physiol 1990;420:431-46.
978
on Novem
ber 20, 2020 by guest. Protected by copyright.
http://thorax.bmj.com
/T
horax: first published as 10.1136/thx.47.11.971 on 1 Novem
