contribution of multiple technique pulmonary medicine* 5 series … · mixed venous po2 at a higher...

Thorax 1994;49:1251-1258

Contribution of multiple inert gas elimination technique topulmonary medicine* 5 Series editor: R Rodriguez-Roisin

Ventilation-perfusion relationships in acuterespiratory failure

C Melot

The optimum approach to therapy in acuterespiratory failure resulting either from theadult respiratory distress syndrome (ARDS) orfrom severe bacterial pneumonia remains achallenge for the clinician. Such therapy isessentially supportive until the basic lung injuryresolves. Effective respiratory support shouldbe provided, taking into account the physio-pathological process underlying the altered pul-monary gas exchange. A precise knowledgeof the mechanisms of arterial hypoxaemia istherefore essential.As mentioned in a previous review in this

series,' standard blood gases and derived in-dices of pulmonary gas exchange such as thevenous admixture (Qva/QT) and the physio-logical dead space (VD/VT) do not permit cleardifferentiation between shunt and units withlow ventilation-perfusion ratios (VA/Q), as wellas between regions having little (high VA/Q)or no perfusion (VA/Q of infinity - that is,anatomical dead space), as the cause of alteredgas exchange. The multiple inert gas elim-ination technique (MIGET)23 takes advantage,not only in recovering the distribution ofVA/Q ratios, but in controlling all the otherfactors determining arterial blood gases in-cluding shunt, VA/Q mismatch, limitation ofdiffusion ofoxygen, and extrapulmonary factorsincluding ventilation, cardiac output, oxygenuptake, haemoglobin, acid-base status, P50 andblood temperature.

However, the MIGET - if it has been amilestone in the comprehensive approach ofthe mechanisms of altered gas exchange inacute respiratory failure - remains to be in-terpreted cautiously when both distributions ofventilation and perfusion are altered sim-ultaneously by a pathological process or bypharmacological and therapeutic interventions.The only way to separate alteration in one ofthese distributions from the other is to performa simultaneous determination by an in-dependent method, either that of the vent-ilation distribution or of the perfusiondistribution.4 Depending on the model used bythe MIGET that constrains all alveoli to have1 of 50 VA/Q values evenly spaced on a log-arithmic scale, the ventilation distribution and

the perfusion distribution are linked together.In order to illustrate the limitation due to theinterdependence between these two dis-tributions in the recovered VA/Q distributionby the MIGET, VA/Q isopleths are plotted ona ventilation versus perfusion diagram (fig 1).For a high VA/Q ratio the slope of the isoplethis steep and a considerable effect on distributionof VA/Q could arise either from subtle changesin blood flow4 or from huge changes in vent-ilation. Conversely, for a low VA/Q the slope isshallow and a change in VA/Q distributioncould result from an easily detectable drop inperfusion or equally from a subtle reduction inventilation (fig 1). In addition, acetone usedby the MIGET to explore high VA/Q areasand dead space is highly soluble in water andconsequently easily lost in the conducting air-ways5 or in the tubing of the ventilator in

cE

0

CL-

m

a)>, 1

0.40.3u

VA/Q= 1

V4/- ---------------- VAQ = 0.1l1,1

0 C?'~00

2 3Blood flow (I/min)

4 5

Figure 1 Alveolar ventilation (VA) versus bloodflow (Q)with VAIQ isopleths (VAIQ = 0.1, 1, and 10). Q andVA are both arbitrarily fixed at 4 1/min. In a low VAIQarea (for example VAIQ = 0.1) the slope is flat and achange in VAIQ distribution could be the result of aneasily measurable decrease in bloodflow from 4 to 3 /lmin(that is, 25% of cardiac output) or equally by a hardlydetectable decrease in VA from 0.4 to 0.3 llmin (that is,2.5% of ventilation). Conversely, in a high VAIQ area(for example, VAIQ = 10) the slope is steep and achange in VAIQ distribution could be the result of aneasily measurable decrease in alveolar ventilation from 4 to3 llmin (that is, 25% of ventilation) or equally by ahardly detectable decrease in QTfrom 0.4 to 0.3 1/min(that is, 2.5% of cardiac ouput).

Intensive CareDepartment, ErasmeUniversity Hospital,Lennik Road 808,B1070 Brussels,Belgium,C Me1otReprint requests to:Dr C Melot.

1251

on July 10, 2020 by guest. Protected by copyright.

http://thorax.bmj.com

/T

horax: first published as 10.1136/thx.49.12.1251 on 1 Decem

ber 1994. Dow

nloaded from

http://thorax.bmj.com/

1252

patients undergoing mechanical ventilation.Changes in high VA/Q areas must thereforebe interpreted cautiously. Nevertheless, despitethese limitations the MIGET is now consideredas a classically reliable technique for the meas-urement of distribution of VA/Q ratios at bed-side in an intensive care setting.

Adult respiratory distress syndrome(ARDS)Although ARDS has been recognised since thefirst description by Ashbaugh et al6 in 1967 asa cause ofacute respiratory failure characterisedby hypoxaemia, the underlying aberrations ingas exchange and their alteration by positiveend expiratory pressure (PEEP) or by phar-macological interventions has been poorly de-fined until studies using the MIGET.

MECHANISMS OF HYPOXAEMIAThe physiological hallmark of ARDS is severehypoxaemia refractory to high concentrationsof inspired oxygen. The pathological basis ofthe hypoxaemia has been shown to be alveolarflooding. The mechanisms of the hypoxaemiahave been variously ascribed, before theMIGET studies, to right-to-left shunt,7 VA/Qinequality,8 and impairment of diffusion.9Moreover, interpretations of altered bloodgases in such patients is complicated by con-comitant physiological stresses on gas ex-change, such as changing cardiac output,anaemia, and acid-base disturbances.The study by Dantzker et al in 1979,10 sub-

sequently confirmed by many others,"-18 was aseminal paper that elucidated with the MIGETthe mechanisms of arterial hypoxaemia inpatients with ARDS. In 16 patients on mech-anical ventilation, eight of whom had had noPEEP, all the patients had increased shunt -that is, perfusion to unventilated lung - from18% to 68% of the cardiac output (fig 2A).Besides shunt, seven patients had low VA/Q

units (fig 2B). All these abnormal VA/Q unitsreceived 48% of blood flow. The remaining52% of blood flow perfused lung units withnormal or increased VA/Q ratios representingeffective gas exchanging units. The presenceof large intrapulmonary shunt explained theprofound hypoxaemia of ARDS poorly re-sponsive to high inspiratory concentrations ofoxygen. The presence of a mode with lowVA/Q units in half of the patients explained theincrease in venous admixture with decreasinginspired fractional concentration of oxygen(FiO2) documented previously by Lamy et al in28 of 45 patients with ARDS.'9 The closeagreement between measured arterial Po2 andpredicted arterial Po2 seen in these patientsargued against a failure ofalveolar-end capillaryequilibrium and ruled out any significant diffu-sion impairment in ARDS.

EFFECTS OF ALTERATION OF CARDIAC OUTPUTSeveral studies, both in humans and in animalmodels, have shown that a change in cardiacoutput resulted in a parallel change in intra-pulmonary shunt. 112123 Several mechanismshave been proposed to explain the observedincrease in shunt with the increase in cardiacoutput: (a) shortened transit time of the eryth-rocytes at a high cardiac output with incompleteequilibrium between alveolar Po2 and end ca-pillary Po2 in these regions;24 (b) an increasedmixed venous Po2 at a higher cardiac outputwith reduced hypoxic pulmonary vaso-constriction and a redistribution of bloodflow preferentially to shunt areas;202526 (c) anincreased amount of oedema in regions of lunginjury at a higher cardiac output.27The effect of changing cardiac output on

intrapulmonary shunt was studied by Lynch eta12' in a canine model of pulmonary oedemainduced by oleic acid. Cardiac output was al-ternately depressed and augmented using eitherpharmacological agents or mechanical al-teration to venous return. Changes in cardiac

A

Pao2 54 mm HgPEEP + 20 cm H2O

Dead spaceShunt 64%33%

-IIA0-01 0.1 1-0 10.0 100.0

VA/Q ratio

1.5

1- .o

0.5 .

0

B

Pao2 64 mm HgPEEP + 16 cm H20

Dead space

,Q 43%

. Shunt13%

C1.5

Pao2 105 mm HgPEEP + 10 cm H20

1.*0

0.5

0

0-01 0-1 1-0 10.0 100*0

VA/Q ratio

Dead space38%

. Shunt14%

------ lw -1ot

.-01

yA/Q ratio

Figure 2 Three main patterns of VA/Q distributions measured using the MIGET in patients with ARDS intubated andmechanically ventilated with PEEP (A) A unimodal normal mode centred on VAIQ = I with an increase in shunt andin dead space. (B) A low VAIQ mode in addition to the normal mode with an increase in shunt and in dead space. (C)A normal VA/Q mode with a high VAIQ mode with an increase in shunt and in dead space. Pao2 = arterial Po2;PEEP = positive end expiratory pressure. Redrawn from references 13 and 14.

[

i.

1.5

._.c

3: 1.00

0

0

0.5c0

0co

Melot



/T


ber 1994. Dow

nloaded from


Ventilation-perfusion relationships in acute respiratory failure

output were not associated with change in theshape of the VA/Q distributions but there wasa significant linear correlation between the levelof cardiac output and the shunt fraction. Theshunt fraction also varied directly with themixed venous Po2. Breen et al,27 comparingoxygen and inert gas exchange, concluded thatincomplete alveolar-end capillary equilibriumfor oxygen contributed very little to any increasein pulmonary shunt with cardiac output inan animal model of pulmonary oedema. Theyhypothesised that cardiac output increasedshunt by increasing oedema or haematocrit inoedematous lung regions.27As in animals with experimentally induced

diffuse lung disease, alterations of cardiac out-put in patients with ARDS have been shownto cause similar adjustments for shunt. Thishas been demonstrated when cardiac outputwas varied by mechanical means with extra-corporeal membrane oxygenation,23 byblood volume expansion,22 and by phar-macological agents.22 The MIGET studies haveshown that, when cardiac output is alteredeither by mechanical" or by pharmacolog-ical means,'31617 the shape of the VA/Qdistributions did not change but shunt in-creased with cardiac output. Sandoval et al25suggested that the increase in shunt, measuredby the Berggren method,28 resulted from anincrease in mixed venous Po2 rather than anincrease in cardiac output itself, suggesting aninhibition of hypoxic pulmonary vaso-constriction in unventilated lung regions.29On the contrary, studies using the MIGETprovided evidence that intrapulmonary shuntincreased with cardiac output even in theabsence of changes in mixed venous Po2(table 1).13 6Moreover, a recent animal study by Domino

et al30 using the MIGET in a lung lobe modelof pulmonary oedema showed that hypoxicpulmonary vasoconstriction was attenuated inthe injured lobe, tempering the role of hypoxicpulmonary vasoconstriction in the observedchanges in shunt.

EFFECTS OF PEEPThe effects of PEEP on VA/Q distributionmainly result from two mechanisms: the effectson shunt and the effects on physiological deadspace.

Table 1 Effects of vasodilators and a vasoconstrictor on pulmonary vascular tone andgas exchange in patients with ARDS

Pap QT Pao2 PVO2 Shunt MIGET studies(Reference no)

Vasodilators:Diltiazem X -X 14Sodium nitroprusside = A 15Ketanserin X = = = 15Nitroglycerin = A 16Prostaglandin E, X~ A X = A 16

X. A X. = A 13Prostacyclin (intravenous) X A = A A 17

X. AX X. = 52Prostacyclin (aerosolised) - A ? 'J 53Nitric oxide -= = X 52

Vasoconstrictor:Almitrine A = A A N. 54

Pap = pulmonary artery mean pressure; QT = cardiac output; Pao, = arterial Po,; Pvo, =mixed venous Po2; shunt = inert gas shunt; A = increased; X. = decreased; = = no change.

Effects on physiological dead spaceThe physiological VDVT calculated from thearterial and mixed expired partial pressuresof carbon dioxide is raised in patients withARDS.'9 Low levels ofPEEP reduce the VDVTin these patients but higher levels consistentlyincrease it.3' Two mechanisms were dem-onstrated by studies using the MIGET: (a)increase in anatomical dead space by dis-tension; (b) derecruitment of some lung unitscreating high VA/Q or dead space units (VA/Qof infinity).The first study in which the MIGET was

used to analyse the effect of PEEP on VA/Qdistributions was by Dueck et al32 in dogs withnormal or oedematous lungs. Increasing PEEPcaused progressive depression ofcardiac outputassociated with an increase in ventilation toboth high VA/Q and unperfused regions. How-ever, at PEEP levels higher than 10 cm H20,retention of carbon dioxide developed exceptin the most severely affected dogs with highestshunt levels. In these there were fewer highVA/Q areas and a further reduction in shuntresulted in a fall in arterial Pco2. Thus, thechanges in arterial Pco2 with increasing PEEP,which differ in moderately and severely affectedanimals, can be explained rationally with theuse of the MIGET. Similarly, Hedenstiema etal' showed that, in the normal dog lung, PEEPcaused a high VA/Q mode with carbon dioxideretention. Moreover, these authors observed aredistribution of pulmonary blood flow withPEEP by measuring blood flow distributionusing an isotopic technique. The study byCoffey et al33 elucidated the effect of PEEPon physiological dead space in dogs with oleicacid-induced pulmonary oedema. Physio-logical dead space can be influenced by changesin anatomical dead space, VA/Q heterogeneity,shunt, and the Haldane effect. Physiologicaldead space decreased with 5 and 10 cm H20PEEP but increased progressively at higherPEEP levels as reported earlier by Suter et al.3'The decrease in physiological dead space at 5or 10 cm H20 PEEP was due to reductions inshunt and mid range VA/Q heterogeneity. Theincrease in physiological dead space that oc-curred with higher PEEP levels was due toincreased ventilation to high VA/Q regions anda larger anatomical dead space. The Haldaneeffect magnified the shunt component of deadspace (as the low oxygen saturation of shuntedblood allows it to carry more carbon dioxideto the pulmonary vein, increasing the Paco2),but reduced the influence of mid range VA/Qheterogeneity.33

Patients with ARDS have a large amount ofventilation distributed to unperfused or poorlyperfused regions. 101218 In a study by Ralph etall8 arterial Pco2 did not change during PEEP,except for a small increase at the highest PEEPlevels. Although the inert gas dead space andventilation to high VA/Q regions did not showconsistent changes, individual patients didshow the appearance of a high VA/Q mode withincreasing PEEP (fig 2C),'0 a response similarto that observed in dogs with oleic acid lunginjury.3233 In patients with ARDS a lesser re-duction in blood flow by PEEP, due to thera-

1253



/T


ber 1994. Dow

nloaded from


Mdlot

peutic intervention attempting to maintaincardiac output, could be the reason for differ-ences between clinical studies'01218 and con-trolled animal studies.3233 Indeed, when cardiacoutput was kept constant during PEEP, a highVA/Q mode did not develop and dead spaceincreased only by 5%.12

Effects on intrapulmonary shuntMost patients showed improvement in arterialP02 with PEEP which mainly resulted from adecrease in the intrapulmonary shunt.'018 Atleast two different mechanisms have been pro-posed to explain the reduction in shunt inducedby PEEP: (a) alveolar re-expansion becauseof the increase in functional residual capacity(FRC) above closing volume,3435 and (b) de-creased perfusion to unventilated lung areasdue to a decrease in cardiac output with PEEP(vascular derecruitment)."Dantzker et al' studied the effect of in-

cremental increases in PEEP to 12 patientswith ARDS on the VA/Q distributions andshowed that PEEP acts by reducing the pro-portion of shunt units that are converted inunits with normal VA/Q by an "on-off" phe-nomenon, indicating that PEEP allows the re-cruitment of non-functional gas exchangingunits. As already discussed, a fall in cardiacoutput results in a parallel decrease in intra-pulmonary shunt. Thus, it has been sug-gested that some of the change observed in theshunt when PEEP is applied may be a resultof the drop in cardiac output."' However, Ma-tamis et al'2 showed that maintenance ofcardiacoutput by dopamine infusion during PEEPventilation did not prevent the expected fall inshunt. The beneficial effect of PEEP on shuntat constant flow was explained by redistributionofblood flow from shunt units to normal units,resulting from alveolar re-expansion rather thanreduction in cardiac output. These results high-light interpretations provided by the MIGETin such a complicated setting in which al-terations of cardiac output by mechanicalmeans were corrected by pharmacologicalagents.

EFFECTS OF BREATHING 100% OXYGENAnother advantage of the MIGET is that itallows the investigator to study the effects ofoxygen breathing on the distribution of VA/Qratios. It has been suggested that inhalationofoxygen increases the quantity ofunventilatedlung regions in normal subjects36 and in patientswith ARDS.37 The VA/Q measurements showedthat, in acute respiratory failure, units with lowVA/Q ratios are converted into shunt.37 Themechanism for this has been related to ab-sorption atelectasis.38 Another possible mech-anism has been the inhibition of hypoxicpulmonary vasoconstriction in shunt units dueto an increase in mixed venous Po2.29

In patients with acute respiratory failure Le-maire et al39 found no difference between intra-pulmonary shunt measured by Berggren'smethod and that of the MIGET. The apparentstability of these alveoli with low VA/Q during

ventilation with oxygen makes it more likelythat they represented fluid-filled alveoli whosevolumes were sufficiently increased at the peakof inspiration to permit some gas transfer.'039Another explanation could be that the largerise in alveolar P02 in the low VA/Q units(less than about 0 1) causes a large increase inoxygen flow into the blood, and this might besufficiently large to exceed the rate of deliveryof inspired gas. Under these conditions suchunits could not eliminate gas and would appearas unventilated by the MIGET. More recentlySantos et al40 showed that, in patients withacute respiratory failure, ventilation with 100%oxygen for one hour increased shunt by 30%.More interestingly, they also showed that shuntremained elevated one hour after returning tothe original FiO2 (below 0 4). The amount ofblood flow to low VA/Q units did not varyduring 100% oxygen, suggesting that increasein shunt was more from absorption atelectasisthan from a release of hypoxic pulmonaryvasoconstriction.

EFFECTS OF BODY POSITION, DIFFERENTIALVENTILATION, AND SELECTIVE PEEPThe large intrapulmonary shunt found inpatients with ARDS requires an increased FIO2and PEEP to minimise hypoxaemia. Both ofthese interventions are associated with wellrecognised complications (oxygen toxicity tothe lung, barotrauma). Some studies showedthat turning patients with ARDS from thesupine to the prone position improved arterialoxygenation.4142 Langer et a143 described 12patients with acute respiratory failure in whomthe prone position caused variable effects ongas exchange. Recently, Gattinoni et al144 con-ducted a prospective study of the effect of asupine to prone change using computed tomo-graphic (CT) scanning of the lungs, andalso measured gas exchange in 10 patients withacute respiratory failure. Whilst changes in thelocation of CT scan densities were observed,the average density was not altered and nosignificant changes were noted in gas exchange.However, large improvements in arterial P02and in intrapulmonary shunt were observed intwo patients. Albert et al45 investigated thisphenomenon using the MIGET in an animalmodel with acute lung injury induced by oleicacid. They showed that the prone positionselectively decreased intrapulmonary shunt andimproved arterial P02. Thus, the prone positionmay be useful as an additional modality toimprove arterial oxygenation in some patientswith ARDS.

In ARDS the matching of ventilation andperfusion in each lung can be improved byapplying differential ventilation with selectivePEEP with the patient lying in a lateral position.The lowered FRC in patients with acute re-spiratory failure may compromise ventilationof dependent lung regions whilst the fractionalperfusion to these regions is increased. PEEPcannot restore the balance between ventilationand perfusion to normal, but this can be sat-isfied further by delivering less gas to non-dependent and more to dependent lung re-

1254



/T


ber 1994. Dow

nloaded from



gions. In practice this can be accomplishedby placing the patient in the lateral position,ventilating each lung separately in proportionto its perfusion (differential ventilation), andapplying PEEP solely to the dependent lung toensure the distribution of gas to the lowerregions of that lung (selective PEEP). By thismeans an improved VA/Q ratio within sep-arately ventilated lungs can be obtained.4"49Interestingly, differential ventilation and se-lective PEEP have been used in patients withbilateral, diffuse lung damage.4"9 MIGETstudies are scarce in this setting but Klingstedtet al50 showed that, in patients without acutelung disease, differential ventilation and se-lective PEEP in the lateral position improvedVA/Q matching with a more even overallVA/Q distribution in both lungs and a decreasedperfusion to shunt and low VA/Q regions.

EFFECTS OF PHARMACOLOGICAL ALTERATION INPULMONARY VASCULAR TONE (table 1)The MIGET has been used in patients withARDS where reduction of pulmonary vasculartone and pulmonary hypertension was obtainedby pharmacological means. Diltiazem, a cal-cium channel blocking agent, reduced pul-monary vascular tone.'4 A reduction inpulmonary vascular tone was obtained withoutchanging either the cardiac output or mixedvenous Po2, with a deterioration in arterialoxygenation. The inert gas data showed a sig-nificant increase in intrapulmonary shunt. Sim-ilar results were reported with sodiumnitroprusside'5 and nitroglycerine.'6 TheMIGETmeasured an increase in blood flow tohypoxic units (with zero and low VA/Q ratios),suggesting a change in gas exchange byinhibition of hypoxic pulmonary vaso-constriction.5' Ketanserin (an antagonist ofserotonin receptors), a pulmonary vasodilatordevoid of any effect on hypoxic pulmonaryvasoconstriction, reduced pulmonary arterypressure in patients with ARDS without de-terioration in gas exchange."

Prostaglandin El (PGEI), a potent shortacting pulmonary vasodilating prostaglandin,

Baseline.c 1.5E

v

o 1.0

00

o 0.5

c

.0

L-

c 0

a,

0.01 0.1 1.0 10.0 100.0VA/Q ratio

caused gas exchange to deteriorate in patientswith ARDS due to a combination of reductionin pulmonary vascular tone and increase incardiac output, without changes in mixed ven-ous Po2.'3'6 VA/Q distributions showed an in-crease in intrapulmonary shunt and a large fallin arterial P02 in some patients (fig 3). To studythe mechanism by which PGE1 increased trueshunt and allowed gas exchange to worsen thelung model used by the MIGET has beenmanipulated so that cardiac output remainedconstant to quantitate the effect of the re-duction in pulmonary vascular tone.'3 De-leterious effects of PGE1 on gas exchangemainly resulted from the reduction in pul-monary vascular tone (table 2) and was presentwith no change in mixed venous Po2. Theseresults must be compared with those obtainedwith intravenous prostacyclin, a pulmonaryvasodilating prostaglandin which caused wor-sening VA/Q distribution in ARDS but not thearterial P02 when mixed venous Po2 increasedsignificantly. 1'752Improvement in gas exchange with reduction

in pulmonary hypertension has recently beenreported with aerosolised prostacyclin in threepatients with severe ARDS.s3 The MIGETshowed that the beneficial effect on gas ex-change resulted from a redistribution of bloodflow from shunt areas to regions of normalVA/Q ratios due to selective pulmonary vaso-dilation in well ventilated areas with aero-solised prostacyclin.

Reyes et al4 reported beneficial effects ongas exchange oftransient perfusion ofalmitrine,a peripheral chemoreceptor agonist with directvasoactive properties on pulmonary vessels.There was an improvement in VA/Q dis-tributions with a reduction in shunt and animprovement in arterial Po2 and mixed venousPo2 at the price ofa slight increase in pulmonaryhypertension. They suggested that almitrinecould enhance hypoxic pulmonary vaso-constriction, diverting blood from shunt unitsto those with normal VA/Q ratios.An exciting further development has been the

identification of nitric oxide as an endotheliumderived relaxing factor55 and the recognition

Prostaglandin E11 *b I

1.0

0.01 0.1 1.0 10.0 100.0VA/Q ratio

Figure 3 Effect ofprostaglandin E, (PGE) 0. 04 jig/kg/min on VA/Q distribution measured using the MIGET in apatient with ARDS intubated and mechanically ventilated with PEEP + 10 cm H20. PGE, increased cardiac outputfrom 6.6 llmin to 7.3 I/min and shunt from 26% to 40%. Pao2 decreasedfrom 95 mmHg to 65 mmHg. QT = cardiacoutput, VE = minute ventilation; Pao2 = arterial Po2. Redrawn from reference 13.

QT 6.6 I/min VE 14.7 I/minPao2 95 mm Hg

j'~~Dead spacel Y l~~53°/

Shunt26%

QT 7.3 I/min VE 14 4 I/min- Pao2 65 mm Hg

~4Shu/Ont Dead spa(

*40%61

0.5 F

0

-F L

. . . . .1255

I

I

ce1%



/T


ber 1994. Dow

nloaded from


Melot

Table 2 Theoretical effect on predicted arterial Po2, calculated from the measured VAIQdistibutions, of increased cardiac output (QT) and reduced pulmonary vascular tone byPGE, in six patients with ARDS. Results are expressed as mean (SE)

PGE,

Baseline Unchanged QT Increased QT

Cardiac index (QT) (1/min/m2) 3 23 (0-26) 3 23 (0.26) 3-86 (0 33)*Pulmonary artery pressure (mmHg) 29 (2) 24 (2)* 24 (2)*Predicted arterial Po2 (mmHg) 97 2 (7-8) 77 8 (8.3)* 75-7 (7.6)*Shunt (% of QT) 20 8 (4 1) 28 4 (5 0)* 31 8 (5 4)*Mixed venous Po2 (mmHg) 36 (2) 38 (3) 38 (3)

* p < 0.01 from baseline value (calculated with data from reference 13).

of a potential role for this agent in matchingventilation and perfusion in acute lung injury.Investigation in vivo is difficult because nitricoxide is the most rapidly binding ligand ofhaemoglobin. Since nitric oxide is inactivatedby haemoglobin, its vasorelaxant effects are

restricted to the vascular smooth muscle under-lying the endothelium. This problem can

be overcome if nitric oxide is inhaled, therebyreaching directly the pulmonary arteriolarsmooth muscles. Rossaint et al showed thatinhalation of nitric oxide causes selective pul-monary vasodilation in ventilated lung regionsand improves VA/Q distributions by divertingblood flow from shunt areas to those of normalVA/Q mode.52 In patients with ARDS theMIGET provided a clear understanding of themechanism of improved gas exchange andarterial Po2, and demonstrated the selectivevasodilation induced by inhaled nitric oxide inventilated areas with normal VA/Q ratios.

EFFECTS OF MIXED VENOUS P02 ON ARTERIAL

OXYGENATIONAs discussed above, the effect ofthe associationbetween cardiac output and shunt on arterialoxygenation will depend on the effect that al-terations in cardiac output have on mixed ven-

ous Po2. In some of the studies in patients withARDS the expected increases in the mixedvenous P02 with increasing cardiac output (ordecreases with decreasing cardiac output) res-

ulted in no change in arterial P02 as shuntvaried (table 1). However, mixed venous P02often failed to behave as expected, sometimesremaining unchanged as cardiac output was

altered and occasionally even decreasing as

cardiac output increased. This was ascribed toa causal relationship between oxygen delivery(Do2) and peripheral oxygen consumption(Vo2). Several studies have identified an ab-normal relationship between Do2 and Vo2 inpatients with ARDS and sepsis.5"6' At rest Vo2is normally independent of Do2 provided thelatter is maintained above a critical level, butoxygen consumption becomes delivery de-pendent above this threshold in ARDS.5"' Themechanisms underlying this observation are

poorly understood, but it has been interpretedas evidence of covert tissue hypoxia and hasbeen associated with a high mortality.5" Anincreased plasma lactate concentation, whichmay reflect an imbalance between metabolicrequirements and Do2, could be a useful marker

of oxygen uptake supply dependency.596062Under these circumstances changes in shuntinduced by cardiac output would have sig-nificant effects on arterial oxygenation.

Severe pneumoniaPatients with pneumonia frequently have ar-terial hypoxaemia. Several factors includingincreased whole body oxygen uptake,63 in-creased intrapulmonary oxygen uptake,64increased intrapulmonary shunt,65 VA/Q mis-matching,65 increased postcapillary shunt (thatis, increased bronchial blood flow),6366 and/oralveolar-end capillary oxygen diffusion lim-itation66 have been implicated as potentialmechanisms of arterial hypoxaemia. The roleof these factors has been clarified by studiesusing the MIGET.

MECHANISMS OF HYPOXAEMIAWagner et al,67 in a canine model of pneumo-coccal lobar pneumonia, showed that pureshunt was present during the first 48 hours ofinfection, whereas after two days the shuntresolved and perfusion was mainly distributedto alveoli with low VA/Q ratios.The study by Lampron et a168 demonstrated

that the most common pattern of VA/Q mis-matching in patients with bacterial pneumoniasevere enough to require mechanical ventilationwas a combination of intrapulmonary shuntand increased perfusion to units with lowVA/Q ratios. The study by Gea et a169 confirmedthese results in patients mechanically ventilatedand extended them to spontaneously breathingpatients with less severe pneumonia. Nodifferences between the predicted and meas-ured arterial oxygen tension were observed inthe MIGET studies, indicating no role foradditional factors such as intrapulmonary oxy-gen consumption, oxygen diffusion limitation,or postcapillary shunt due to increased bron-chial blood flow.

EFFECTS OF BREATHING 100% OXYGENReports ofthe effects of 100% oxygen breathingon VA/Q relationships in patients with pneu-monia are conflicting. It had been reported thatshunt may remain unchanged or may evenincrease when the patient is breathing 100%oxygen. Since a substantial proportion of pul-monary blood flow is diverted to low VA/Qunits, the development ofabsorption atelectasisis expected so that shunt should increase inthese patients.Lampron et al68 were unable to find an in-

crease in intrapulmonary shunt with theMIGET in patients with severe pneumoniaduring ventilation with 100% oxygen. Similarresults were also reported by Gea et al69 whoshowed a widening of the blood flow dis-tribution, suggesting a reduction in hypoxicpulmonary vasoconstriction. However, neitherstudy showed an increase in shunt despite thefact that mixed venous Po2 increased, sug-gesting only minimal hypoxic pulmonary vaso-constriction in human bacterial pneumonia.

1256



/T


ber 1994. Dow

nloaded from



EFFECTS OF BODY POSITION

In 1981 Fishman70proposed the lateral positionfor patients with single lung injuries. Severalinvestigators reported a better oxygenation inpatients in the lateral position with the goodlung dependent.7"- The likely mechanism ofimproved gas exchange may be better VA/Qmatching; perfusion being gravity dependent isdirected towards the good lung, while only a

small fraction of pulmonary blood flow but a

greater portion of tidal volume is distributedto the non-dependent lung. Gillespie et al74reported four patients with unilateral diseasewho improved their arterial oxygenation whenthe good lung was dependent. The MIGEThas confirmed that positional changes cause

reduction in shunt or in areas with low VA/Qratios or in both.74

ConclusionsThe MIGET has made it possible to differ-entiate changes in the arterial Po2 caused byalterations in the gas exchanging function ofthe lung from those due to changes in themixed venous P02. It has also played a part inassessing the effects of alterations in cardiacoutput, pulmonary vascular tone, or both on

arterial blood gas tensions in patients with acuterespiratory failure.

1 Roca J, Wagner PD. Principles and information content of

the multiple inert gas elimination technique. Thorax 1994;49:815-24.

2 Wagner PD, Saltzman HA, West JB. Measurement of con-

tinuous distributions of ventilation-perfusion ratios:

theory. Appl Physiol 1974;36:588-99.3 Wagner PD, Naumann PF, Laravuso RB. Simultaneous

measurement of eight foreign gases in blood by gas chro-

matography. Appl Physiol 1974;36:600-5.4 Hedenstierna G, White FC, Mazzone R, Wagner PD. Re-

distribution of pulmonary blood flow in the dog withPEEP ventilation. Appl Physiol 1979;46:278-87.

5 SchrikkerAC, de-Vries WR, ZwartA, Luijendijk SC. Uptakeof high soluble gases in the epithelium of the conductingairways. Pflugers' Arch 1985;405:389-94.

6 Ashbaugh DG, Bigelow DB, Petty TL, Levine BE. Acute

respiratory distress in adults. Lancet 1967;ii:319-23.7 Pontoppidan H, Geffin B, Lowenstein E. Acute respiratory

failure in the adult. NEnglJMed 1972;287:690-8,743-52,799-806.

8 Markello R, Winter P, Olszowka A. Assessment of vent-

ilation-perfusion inequalities by arterial-alveolar nitrogendifferences in intensive-care patients. Anesthesioloogy 1972;37:4-15.

9 King TK, Weber B, Okinaka A, Friedman SA, Smith JP,Briscoe WA. Oxygen transfer in catastrophic respiratoryfailure. Chest 1974;65(Suppl):40-4S.

10 Dantzker DR, Brook CJ, DeHart P, Lynch JP, Weg JG.Ventilation-perfusion distributions in the adult respiratorydistress syndrome. Am Rev RespirDis 1979;120:1039-52.

11 Dantzker DR, Lynch JP, Weg JG. Depression of cardiac

output is a mechanism of shunt reduction in the therapyof acute respiratory failure. Chest 1980;77:636-42.

12 Matamis D, Lemaire F, Harf A, Teisseire B, Brun-BuissonC. Redistribution of pulmonary blood flow induced bypositive end-expiratory pressure and dopamine infusionin acute respiratory failure. Am Rev Respir Dis 1984;129:39-44.

13 Melot C, Lejeune P, Leeman M, Moraine JJ, Naeije R.Prostaglandin E, in the adult respiratory distress syn-drome. Benefit for pulmonary hypertension and cost forpulmonary gas exchange. Am Rev Respir Dis 1989;139:106-10.

14 Melot C, Naeije R, Mols P, Hallemans R, Lejeune P, JasparN. Pulmonary vascular tone improves pulmonary gasexchange in the adult respiratory distress syndrome. AmRev Respir Dis 1987;136:1232-6.

15 Radermacher P, Huet Y, Pluskwa F, Herigault R, Mal H,Teisseire B, et al. Comparison of ketanserin and sodiumnitroprusside in patients with severe ARDS. Anesthesiology1 988;68: 152-7.

16 Radermacher P, Santak B, Becker H, Falke KJ. Pro-staglandin E, and nitroglycerin reduce pulmonary capillarypressure but worsen ventilation-perfusion distributions inpatients with adult respiratory distress syndrome.Anesthesiology 1989;70:601-6.

17 Radermacher P, Santak B, Wust HJ, Tamow J, Falke KJ.Prostacyclin for the treatment of pulmonary hypertensionin the adult respiratory distress syndrome: effects on pul-

monary capillary pressure and ventilation-perfusion dis-tributions. Anesthesiology 1990;72:238-44.

18 Ralph DD, Robertson, HT, Weaver U, Hlastala MP, Carrico

CJ, Hudson LD. Distribution of ventilation and perfusionduring positive end-expiratory pressure in the adult re-

spiratory distress syndrome. Am Rev Respir Dis 1985;131:54-60.

19 Lamy M, Fallat RJ, Koeniger E, Dietrich HP, Ratliff JL,Eberhart RC, et al. Pathologic features and mechanisms

of hypoxemia in adult respiratory distress syndrome. AmRev Respir Dis 1976;114:267-84.

20 Smith G, Cheney FWJ, Winter PM. The effect of change in

cardiac output on intrapulmonary shunting. BrJAnaesth1974:46:337-42.

21 Lynch HP, Mhyre JG, Dantzker DR. Influence of cardiac

output on intrapulmonary shunt. ApplPhysiol 1979;46:315-21.

22 Jardin F, Eveleigh MC, Gurdjian F, Delille F, Margairaz A.

Venous admixture in human septic shock: comparativeeffects of blood volume expansion, dopamine infusion and

isoproterenol infusion or mismatching of ventilation and

pulmonary blood flow in peritonitis. Circulation 1979;60:155-9.

23 Lemaire F, Jardin F, Regnier B, Loisance D, Goudot B,Lange F, et al. Pulmonary gas exchange during venoarterial

bypass with a membrane lung for acute respiratory failure.

Thorac Cardiovasc Surg 1978;75:839-46.24 Wagner PD. Diffusion and chemical reaction in pulmonary

gas exchange. Physiol Rev 1977;57:257-312.25 Sandoval J, Long GR, Skoog C, Wood LD, Oppenheimer

L. Independent influence of blood flow rate and mixed

venous Po2 on shunt fraction. Appl Physiol 1983;55:1128-33.

26 Suter PM. Fairley HB, Schlobohm RM. Shunt, lung volumeand perfusion during short periods of ventilation with

oxygen. Anesthesiology 1975;43:617-27.27 Breen PH, Schumacker PT, Hedenstierna G, Ali J, Wagner

PD, Wood LD. How does increased cardiac output in-

crease shunt in pulmonary edema? Appl Physiol 1982;53:1273-80.

28 Berggren SM. The oxygen deficit of arterial blood caused

by non-ventilating parts of the lung. Acta Physiol Scand

1942;4(Suppl 11):4-92.29 Domino KB, Wetstein L, Glasser SA, Lindgren L, Marshall

C, Harken A, et al. Influence of mixed venous oxygentension (Po2) on blood flow to atelectatic lung.Anesthesiology 1983;59:428-34.

30 Domino KB, Cheney FW, Eisenstein BL, Hlastala MP.

Effect of regional alveolar hypoxia on gas exchange in

pulmonary edema. Am Rev Respir Dis 1992;145:340-7.31 Suter PM, Fairley B, Isenberg MD. Optimum end-ex-

piratory airway pressure in patients with acute pulmonaryfailure. NEnglJtMed 1975;292:284-9.

32 Dueck R, Wagner PD, West JB. Effects of positive end-

expiratory pressure on gas exchange in dogs with normal

and edematous lungs. Anesthesiology 1977;47:359-66.33 Coffey RL, Albert RK, Robertson HT. Mechanisms of

physiological dead space response to PEEP after acute

oleic acid lung injury. JfAppl Physiol 1983;55:1550-7.34 Falke KJ, Pontoppidan H, Kumar A, Leith DE, Geffin B,

Laver MB. Ventilation with end-expiratory pressure in

acute lung disease. Clin Invest 1972;51:2315-23.35 Ramachandran PR, Fairley HB. Changes in functional re-

sidual capacity during respiratory failure. Can Anaesth Soc

1970;17:359-69.36 Wagner PD, Laravuso RB, Uhl RR, West JB. Continuous

distributions of ventilation-perfusion ratios in normal sub-

jects breathing air and 100 per cent°2-.J Clin Invest 1974;54:54-68.

37 West JB, Wagner PD. Pulmonary gas exchange. In: West

JB, ed. Bioengineering aspects of the lung. New York: Marcel

Dekker, 1977:361-457.38 Dantzker DR, Wagner PD, West JB. Instability of lung units

with low VA/Q ratios during 02 breathing. Appl Physiol1975;38:886-95.

39 Lemaire F, Matamis D, Lampron N, Teisseire B, Harf A.

Intrapulmonary shunt is not increased by 100% oxygenventilation in acute respiratory failure. Bull Eur Physio-pathol Respir 1985;21:251-6.

40 Santos C, Roca J, Torres A, Cardus J, Barbera JA, Ro-

driguez-Roisin R. Patients with acute respiratory failure

increase shunt during 100% 02 breathing. Eur Respir1992;5(Suppl 15):272s.

41 Douglas WW, Rehder K, Beynen FM, Sessler AD, Marsh

HM. Improved oxygenation in patients with acute re-

spiratory failure: the prone position. Am Rev Respir Dis

1977;115:559-66.42 Piehl MA, Brown RS. Use of extreme position changes in

acute respiratory failure. Crit Care Med 1976;4:13-4.43 Langer M, Mascheroni D, Marcolin R, Gattinoni L. The

prone position in ARDS patients. A clinical study. Chest

1988;94: 103-7.44 Gattinoni L. Pelosi P, Vitale G, Pesenti A, D'Andrea L,

Mascheroni D. Body position changes redistribute lungcomputed-tomographic density in patients with acute re-

spiratory failure. Anesthesiology 1991;74:15-23.45 Albert RK, Leasa D, Sanderson M, Robertson HT, Hlastala

MP. The prone position improves arterial oxygenationand reduces shunt in oleic-acid-induced acute lung injury.Am Rev Respir Dir 1987;135:628-33.

46 Hedenstierna G, Baehrendtz S, Klingstedt C, Santesson J,Soderborg B, Dahlborn M, et al. Ventilation and perfusionof each lung during differential ventilation with selectivePEEP. Anesthesiology 1984;61:369-76.

1257



/T


ber 1994. Dow

nloaded from


Melot

47 Baehrendtz S, Hedenstierna G. Differential ventilation andselective positive end-expiratory pressure: effects onpatients with acute bilateral lung disease. Anesthesiology1984:61:511-7.

48 Baehrendtz S, Bindslev L, Hedenstiema G, Santesson J.Selective PEEP in acute bilateral lung disease. Effect on

patients in the lateral posture. Acta Anaesthesiol Scand1983;27:311-7.

49 Baehrendtz S, Santesson J, Bindslev L, Hedenstiema G,Matell G. Differential ventilation in acute bilateral lungdisease. Influence on gas exchange and central haemo-dynamics. Acta Anaesthesiol Scand 1983;27:270-7.

50 Klingstedt C, Hedenstierna G, Baehrendtz S, LundqvistH, Strandberg A, Tokics L, et al. Ventilation-perfusionrelationships and atelectasis formation in the supine andlateral positions during conventional mechanical anddifferential ventilation. Acta Anaesthesiol Scand 1990;34:421-9.

51 Naeije R, Melot C, Mols P, Hallemans R. Effects of vaso-dilators on hypoxic pulmonary vasoconstriction in normalman. Chest 1982;82:404-10

52 Rossaint R, Falke KJ, Lopez F, Slama K, Pison U, ZapolWM. Inhaled nitric oxide for the adult respiratory distresssyndrome. N Engl J7 Med 1993;328:399-405.

53 Walmrath D, Schneider T, Pilch J, Grimminger F, SeegerW. Aerosolised prostacyclin in adult respiratory distresssyndrome. Lancet 1993;342:961-2.

54 Reyes A, Roca J, Rodriguez-Roisin R, Torres A, Ussetti P,Wagner PD. Effect of almitrine on ventilation-perfusiondistribution in adult respiratory distress syndrome. AmRev Respir Dis 1988;137:1062-7.

55 Palmer RM, Ferrige AG, Moncada S. Nitric oxide releaseaccounts for the biological activity of endothelium-derivedrelaxing factor. Nature 1987;327:524-6.

56 DanekSJ, LynchJP, WegJG, Dantzker DR. The dependenceof oxygen uptake on oxygen delivery in the adult re-

spiratory distress syndrome. Am Rev Resp-ir Dis 1980;122:387-95.

57 Mohsenifar Z, Goldbach P, Tashkin DP, Campisi DJ. Re-lationship between O2 delivery and O2 consumption inthe adjult respiratory distress syndrome. Chest 1983;84:267-71.

58 Bihari D, Smithies M, Gimson A, Tinker J. The effects ofvasodilation with prostacyclin on oxygen delivery anduptake in critically ill patients. N Engl Med 1987;317:397-403.

59 Fenwick JC, Dodek PM, Ronco JJ, Phang PT, Wiggs B,Russel JA. Increased concentrations of plasma lactatepredict pathologic dependence of oxygen consumption on

oxygen delivery in patients with adult respiratory distresssyndrome. Cnit Care 1990;5:81-7.

60 Vincent JL, Roman A, De-Backer D, Kahn RJ. Oxygen

uptake/supply dependency. Effects of short-term do-butamine infusion. Am Rev Respir Dis 1990;142:2-7.

61 Clarke C, Edwards JD, Nightingale P, Mortimer AJ, MorrisJ. Persistence of supply dependency of oxygen uptake athigh levels of delivery in adult respiratory distress syn-drome. Crit Care Med 199 1;19:497-502.

62 Annat G, Viale JP, Percival C, Froment M, Motin J. Oxygendelivery and uptake in the adult respiratory distress syn-drome. Lack ofrelationship when measured independentlyin patients with normal blood lactate concentrations. AmRev Respir Dis 1986;133:999-1001.

63 Davidson FF, Glazier JB, Murray JF. The components ofthe alveolar-arterial oxygen tension difference in normalsubjects and in patients with pneumonia and obstructivelung disease. Am Med 1972;52:754-62.

64 Light RB. Intrapulmonary oxygen consumption in ex-

perimental pneumococcal pneumonia. Appl Physiol1988;64:2490-5.

65 Goldzimer EL, Wagner PD, Moser KM. Sequence of vent-ilation/perfusion alterations during experimentalpneumococcal pneumonia in dogs. Chest 1973;64:394-5(abstract).

66 Alexander JK, Takezawa H, Abu-Nassar HJ, York EM.Studies on pulmonary blood flow in pneumococcal pneu-monia. Cardiovasc Res Cent Bull 1963;1:86-92.

67 Wagner PD, Laravuso RB, Goldzimmer E, Naumann PF,West JB. Distribution of ventilation-perfusion ratios indogs with normal and abnormal lungs. Appl Physiol1975;38:1099-109.

68 Lampron N, Lemaire F, Teisseire B, Harf A, Palot M,Matamis D, et al. Mechanical ventilation with 100%oxygen does not increase intrapulmonary shunt in patientswith severe bacterial pneumonia. Am Rev Respir Dis 1985;131:409-13.

69 Gea J, Roca J, Torres A, Agusti AGN, Wagner PD, Ro-driguez-Roisin R. Mechanisms of abnormal gas exchangein patients with pneumonia. Anesthesiology 1991;75:782-9.

70 Fishman AP. Down with the good lung (editorial). N EnglJ'Med 1981;304:537-8.

71 Zack MB, Pontoppidan H, Kazemi H. The effect of lateralpositions on gas exchange in pulmonary disease. A pro-spective evaluation. Am Rev Respir Dis 1974;110:49-55.

72 Remolina C, Khan AU, Santiago TV, Edelman NH. Po-sitional hypoxemia in unilateral lung disease. N EnglMed 1981;304:523-5.

73 Dreyfuss D, Djedaini K, Lanore JJ, Mier L, Froidevaux R,Coste F. A comparative study of the effects of almitrinebimesylate and lateral position during unilateral bacterialpneumonia with severe hypoxemia. Am Rev Respir Dis1992;146:295-9.

74 Gillespie DJ, Rehder K. Body position and ventilation-perfusion relationships in unilateral pulmonary disease.Chest 1987;91:75-9.

1258



/T


ber 1994. Dow

nloaded from


contribution of multiple technique pulmonary medicine* 5 series … · mixed venous po2 at a higher...

Documents