journal, regulation ofthe pulmonary circulation

British Heart Journal, I971, 33, Supplement, I5-26.

Regulation of the pulmonary circulation

G. de J. LeeFrom the Cardiac Department, Radcliffe Infirmary, Oxford

Factors regulating pressure andflow in the lungs are reviewed with particular emphasis on theirrole in regulating blood flow velocity and distribution within the lung capillaries. Thebehaviour of the pulmonary arterial system, alveolar capillaries, and pulmonary venoussystem are considered individually. The effect of heart disease on lung capillary blood flow isexamined.

The behaviour of the lung circulation hasinspired intense study during the years sincethe International Society of Cardiology lastmet in I966. This is partly owing to the factthat techniques are now available for use inman which give information with a precisionpreviously available only to the investigator inthe experimental laboratory. Thus contribu-tions which clinicians have made to the betterunderstanding of the physiological relation-ships governing both ventilation and bloodflow in the lungs have also led to greateraccuracy in diagnosis and a more rationalapproach to treatment.The prime function of the lungs is to

accomplish efficient gas exchange between theblood flowing through the pulmonary capil-laries and the air ventilating their relatedalveoli. Normally this process takes place overan enormous range of activity with the cardiacoutput approaching 30 1./min. in extremeexercise. Yet the lung capillary blood volumeprobably never exceeds 200 ml. (Roughtonand Forster, I957; Bates et al., I960; Weibel,I963), and blood flow is accomplished in sucha manner that the intravascular pressureswithin the capillaries rarely exceed the plasmaosmotic filtration pressure, so that the Starlingrelation between the vascular and extra-vascular compartments of the lungs is main-tained and pulmonary oedema is avoided.

In disease ofthe lungs, adaptive mechanismsoperate which redistribute both blood flowand gas delivery away from diseased areas tothe more normal parts of the lung; while inheart disease, particularly those conditionsleading to increased impedance to venous out-flow from the lungs into the left side of theheart, mechanisms again develop which partlyprotect the lung capillaries from experiencingthe high vascular pressures present in the

larger blood vessels as a consequence of theheart disease. This assists in preserving thealveolar capillary system as a gas exchangingarea rather than converting it to a plasma fil-tration system, manifest clinically by pulmon-ary oedema. The activity ofthe lung lymphaticsystem in draining the extravascular spaces ofthe lung also assists in preserving the gaseousenvironment of the alveoli.The peripheral lung blood vessels also act

as a sieve and as a lytic environment capable ofremoving particulate material such as plateletaggregates, etc., delivered to it from the sys-temic venous system, which would be danger-ous to the organism if allowed to continue tocirculate to the systemic arterial system.

Finally, the systemic circulation contributesto the overall regulation of blood flow in thelungs via the bronchial circulation. In healththe bronchial circulation probably contributesless than I2 per cent of the total lung bloodflow, but in lung disease bronchopulmonaryanastomoses play an increasingly importantpart in local perfusion of diseased areas(Cudkowicz, I968).This brief description of functions ignores

the more occult participation of the lungcirculation in the organ's metabolic andhumoral activity with regard to such sub-stances as histamine, 5-hydroxy tryptamine,and other circulating catecholamines, lipidand cholesterol metabolism, etc. (Heinemannand Fishman, I969).

This brief review of functions involving thelung circulation indicates its complexities.Not unnaturally, the regulation ofthe pulmon-ary circulation depends on an equally complexvariety of factors whose relationships to oneanother result in overall homoeostasis.

Fig. i lists these interrelationships in dia-gramatic form. The 'active' vasomotor effects

on February 17, 2022 by guest. P

rotected by copyright.http://heart.bm

j.com/

Br H

eart J: first published as 10.1136/hrt.33.Suppl.15 on 1 January 1971. D

ownloaded from

http://heart.bmj.com/

x6 G. deJI. Lee

LUNG MECHANICS:

lung volume effectson interstitial

pressure and vascularconfiguration

respiratory activity:ettect on venous

return

I*

Active effectsPassive effects

LEFT ATRIALPRESSURE

Irole of LA andpulmonaryvenous

compliance

I

INSPIRED GASES

hypoxiahypercopnoea

CATECHOLAMINES

VAGAL AND SYMPATHETIC NERVE ACTIVITY

systemic systemic cardiomvKsomotor vasomotor

alveolar gases -*blood -_systemic bronchiclblood vessels

IOPULMONARY VASCULAR RESPONSE

notor bronchomotor PULMONARYVASOMOTOR

t

FIG. i Active and passive factors affecting the pulmonary vascular bed.

referred to in the figure indicate responseswhich occur in the pulmonary blood vesselsmanifest by contraction or relaxation of theirown walls due to humoral, chemical, or reflexstimuli; while 'passive' effects refer to changesimposed upon the pulmonary blood vesselsfrom mechanical events within the thorax andlungs, or secondary to circulatory eventstaking place elsewhere in the body.A detailed analysis of the regulatory func-

tions of each of the activities listed in Fig. i isbeyond the scope of this presentation. Itherefore intend to confine myself to anexamination of what I believe to be the mostimportant factors responsible for regulatingblood flow within the lung capillaries them-selves: for the ultimate test of the lung'sefficiency must surely be its capability tomaintain adequate conditions for gas exchangeover the widest possible range of physiologicalactivity and pathological insult.

Nature of lung capillary blood flowThe lung circulation in normal man has a lowperipheral resistance and a low inherent reflexvasomotor activity (Lee, Matthews, andSharpey-Schafer, 1954). Thus one mightexpect that the pulsatile ejection of blood flowfrom the right ventricle might continue to betransmitted through the pulmonary arterialsystem so that capillary blood flow also re-mained pulsatile. Lee and DuBois (I955)proved this to be the case by measuring thelung capillary blood flow instantaneouslythroughout the cardiac cycle by continuouslyrecording the rate of nitrous oxide uptakefrom the lungs using the body-plethysmo-graph method. They found that the lungcapillary blood flow was highly pulsatile witheach cardiac cycle. These events were so rapidthat some passive rather than reflex vascular

adjustment could be presumed to exist forregulating flow and pressure. It could also beinferred from the fact that the fully distendedlung capillary system has a blood volumealmost identical to the stroke volume of theheart, while many previous workers haddemonstrated that pulmonary vascular resis-tance always decreased when blood flowthrough the lungs increased during exercise.

Blood flow distribution in lungs: ventila-tion-perfusion relationshipsThe hydrostatic mechanism which operates toreduce lung vascular resistance as blood flowincreases also serves to regulate distribution ofblood flow within the lungs. Its existence wasfirst discovered by Banister and Torrance inI960 in isolated perfused lungs. They foundthat the interrelation between alveolar pres-sure, tending to close lung capillaries, and thearterial and venous pressures, tending to openthem, regulated blood flow. The alveolarcapillary systems were acting effectively asStarling resistors.In the same year West and Dollery (I960),

using radioactive gases, had found in man thatblood flow at the lung bases was some eighttimes greater than at the apices, while the dis-tribution of ventilation was also somewhatgreater to the lower zones than it was to theupper zones of the lungs. Intense study oflung ventilation-perfusion distribution fol-lowed, led particularly by Permutt, Brom-berger-Barmea, and Bane (I962) in theUnited States, and West, Dollery, andNaimark (I964) in Britain.

In the vertically suspended isolated, venti-lated, and perfused lung it was found that nocapillary blood flow took place at hydrostaticlevels above which alveolar pressure exceededboth the pulmonary arterial and venous pres-



j.com/

Br H


ownloaded from


Regulation of pulmonary circulation 17

sures (West: Zone i). Flow began at a leveldown the lung where pulmonary arterial pres-sure exceeded alveolar pressure (West: Zone2). It continued to rise in a manner deter-mined by the difference between pulmonaryarterial and alveolar pressures until a level wasreached where the pulmonary venous pressurealso exceeded the alveolar pressure. Over theremainder of the dependent lung blood flowwas then governed by the pulmonary arterial-venous pressure difference (West: Zone 3),for in this zone both these pressures exceededthe alveolar pressure and the capillary systemswere thus permanently open.Hughes et al. (I968) subsequently extended

this work to discover the effects of changes inlung volume on the regional distribution ofblood flow and found that vessels not directlyexposed to alveolar pressure affected vascularresistance in the lowermost parts of the lungin an unexpected way. These so-called extra-alveolar vessels of the lung are held open bythe degree of expansion of the lung (Rosen-zweig, Hughes, and Glazier, I970). However,the base ofthe upright lung is relatively poorlyexpanded, so that the extra-alveolar vesselssignificantly reduce blood flow in the lower-most zones. Hughes and his colleagues wereable to show a fourth zone at the extreme baseof the lung, where regional blood flow beganto fall again in spite of a continuing rise inintravascular hydrostatic pressure.

In the past, Howell et al. (I96I) and Permutt(I965) had produced evidence that lung infla-tion increased the capacity of those larger lungvessels outside the influence of alveolar pres-sure. These vessels were in effect sucked uponby the negative interstitial pressure aboutthem. Mead, Takishima, and Leith (I967)suggested that the negative distending pres-sure transmitted to these vessels depended onthe extent to which they failed to take part inthe expansion process of the lung in inspira-tion. The interstitial tissues in these areaswould then tend to have disproportionatelyhigher negative pressures about them, so thatinterstitial fluid accumulation would tend tobe greater at the base of the lung than in theupper zones, thus contributing to the in-creased vascular resistance to blood flowfound in Zone 4 by Hughes and his colleagues.This has important clinical relevance, for itseems most probable that the increased pul-monary venous pressure which results fromleft-sided heart disease will increase fluidtransudation at the lung bases long beforeovert pulmonary oedema is recognized. Thiswill further increase impedance to blood flowthrough the lower zones of the lungs and willresult in the preferential redistribution of

blood flow to the upper zones of the lungs socommonly seen in mitral stenosis.These hydrostatic relationships describe

blood flow distribution in the lungs understeady flow conditions. They plainly dependon the role of alveolar capillary systems actingas parallel resistors stacked vertically, andindicate how vascular pressure adjustments aswell as zonal distribution of flow could beachieved under real life conditions of pulsatileblood flow. Thus at peak capillary flow ratethe pulmonary arterial pressure will moment-arily exceed the alveolar gas pressures in allareas of the lungs, so that all lung capillarysystems will open to accommodate blood flow.During diastole, alveolar capillary systems willcease to conduct the flow as the pulmonaryarterial pressure falls in their zone. This willbegin in the uppermost zones of the lung andextend downwards as arterial pressure fallsduring diastole. Thus regional lung perfusionwill vary with time throughout the cardiaccycle in a tidal manner, regulated by gravityand becoming most uniform momentarilyduring the peak of systole, or when exerciseincreases the rate of blood flow so that thepulmonary arterial pressure is permanentlyelevated above alveolar pressure.This tidal concept for accommodating pul-

satile capillary blood flow implies a fluctuatingcapillary volume participating in gas diffusionat low levels of cardiac output, while at highoutput levels, when all capillaries are per-manently recruited, increased gas diffusionwill be achieved by pulsatile changes in bloodvelocity through each capillary system.Du Bois and his colleagues (Menkes et al.,

I970) have now demonstrated that the lungcapillary volume does indeed fluctuatethroughout the cardiac cycle in normal erectsubjects at rest. They measured this by an in-genious modification of the single breath Dcomethod, using a water-filled body plethysmo-graph. They showed that the CO uptake fromfive normal subjects indicated large capillaryvolume changes during each cardiac cycle.The greatest volume increase during thecycle coincided with peak capillary bloodflow rates measured by the N20 uptakemethod in the same subject.Lung ventilation is also found to be in-

homogeneous, for it also increases down thelung but less rapidly than does blood flowunder normal conditions. Milic-Emili and hiscolleagues (I966) have shown that subjectsinspiring after full exhalation filled the upperzones of their lungs first. When a certain lungvolume was reached the lower zones began tofill better than the upper zones. This is due tothe way the lungs are supported in the thorax.



j.com/

Br H


ownloaded from


iS G. deJ. Lee

It appears that the hilum takes very little ofthe weight of the lung, which is largely sup-ported by the pleural surfaces consequentupon the negative intrathoracic pressure. Be-cause of gravity, the lower regions are rela-tively compressed by the weight of the lungabove them, and the upper zones are relativelyoverexpanded by the weight ofthe lung below.These differences can be shown by the factthat the intrapleural pressure is less negativeat the bottom of the upright lung than at thetop ofit. A further consequence ofthis gravita-tional effect is the fact that the lung alveolidecrease in size from apex to base in the up-right lung (Glazier et al., I966, I968).

Haemodynamic factors affecting pulsa-tile capillary blood flowHaving briefly reviewed the mechanical effectsthat the physical environment within thethorax imposes upon the lung circulation andthe distribution of pulmonary capillary bloodflow in particular, I now wish to examine theeffect that changes in the intrinsic behaviourof the pulmonary circulation itself may haveon lung capillary blood flow. The pulsatilenature of lung capillary blood flow dependsupon the instantaneous pressure events takingplace simultaneously at the arteriolar andvenular ends of the capillary system. These inturn depend upon simultaneous events in theright ventricle and left atrium of the heart,modified by the physical characteristics of thepulmonary arterial and venous systems linkingthese chambers to the capillary system.The nitrous oxide/body plethysmograph

method for measuring the lung capillaryblood flow, combined with simultaneousmeasurements of vascular pressures, has pro-vided special opportunity to examine theeffects of changes in the pulmonary arterialand venous system upon lung capillary bloodflow pulsatility.We have developed a recording system for

use with the whole body plethysmographwhich allows us to measure the rate of N20uptake from the lungs with a frequency re-sponse that is flat to IS cycles per second(Bosman et al., I964; Karatzas, Lee, and Stott,I967), so that we are now able to comparedynamic flow events with the same accuracyas simultaneous pressure events within thelung circulation and submit both to Fourieranalysis when required.

Fig. 2 shows an example of an N20 up-take record superimposed upon a previouslyobtained air control record from a normalsubject lying in the body plethysmograph.The records are timed against the electro-cardiograph and phonocardiogram in order to

1 sec.

FIG. 2 Parts of an air control record (AIR)and N20uptake record (N20) obtained withidentical heart rates, superimposed upon oneanother.The electrocardiogram and phonocardiogram(PCG) are displayed for timing purposes.The shaded area on the left corresponds to thevolume of N20 absorbed in one cardiac cycle;this is also given by the integrator (SV).DV=N20 uptake during diastole.Ct =Flow conduction time.Ut=N20 upstroke time.Qc = Capillary flow rate cakulated from

N20-air records and redrawn.

time the opening and closing of the pulmonaryvalve (S1, S2, respectively). The shaded area onthe left of the figure indicates the volume ofN20 taken up by the lungs in one cardiaccycle, which is also given by the integratorsignal (SV). On the right of the figure isplotted the absolute capillary blood flow rate(Qc) obtained from the shaded N20 curve.The average acceleration of blood flow in thelung capillaries can be measured from the up-stroke time (Ut) of the N20 curve.

In addition, the percentage of the rightventricular stroke volume stored in the pul-monary arterial system during systole can becalculated. This pulmonary arterial systolicstorage volume can be measured by subtractingthe volume of blood flowing from the lungcapillaries during diastole from the totalstroke volume (stroke volume minus diastolicvolume: SV-DV).The degree of pulmonary capillary flow

pulsatility can be quantitated as a ratio be-tween the peak flow rate to mean flow rate(Qc max/Qc). This is termed the pulsatilityindex.The method also allows one to measure the



j.com/

Br H


ownloaded from


Regulation of pulmonary circulation I9

time it takes the flow pulse to travel from thepulmonary valve to the lung capillaries. Thisis timed either from the first upstroke of thepulmonary artery pressure measured at car-diac catheterization to the time of beginningofN2O uptake in the body plethysmograph, orit may be timed from the first heart sound ofthe phonocardiogram (Sl) to the foot of thecapillary flow pulse measured by N20 uptake.This is termed the pulmonary arterial conduc-tion time (Ct). Our findings from normal sub-jects have shown that the lung capillarypulsatility index (Qc max/Qc) has an approxi-mate ratio of 2:I (Karatzas and Lee, I966).This value agreed closely with similar studiesreported by Wasserman, Butler, and VanKessel (I966). The normal PA-Pc conductiontime averaged I20 ± 7 0 msec., while the sys-tolic acceleration rate (Ut) was 8'2 ± I7ml./sec./msec. (Karatzas, and Lee, I969).Of particular interest was the finding that

the systolic storage volume of the pulmonaryarterial system in normal resting subjectswhose heart rate averaged 6i/min. averaged66 per cent of the total stroke volume. Whenthe heart rate was increased by administeringatropine there was a highly significant fall inthe storage capacity of the pulmonary arterialsystem, so that at an average heart rate ofi io/min. the storage volume had decreased to56 per cent of the stroke volume (Karatzas etal., I969). This finding is really an indicationof the supply capacity of the pulmonaryarterial system available for maintainingefficiency of gas exchange in the lung capil-laries during exercise. Thus during exercise,when the heart rate increases, the diastolicperiod of the cardiac cycle shortens so thatblood flow is presented to the lung capillariesfor gas exchange in a more uniform mannerwith respect to time than under resting con-ditions. This, coupled with the more uniformspatial distribution of blood flow within thelungs during exercise, compared with theresting state, explains the facility with whichthe lung is able to increase its gas exchangecapacity over a wide range of exercise.Our early clinical comparisons of simul-

taneous puilmonary vascular pressure andpulmonary capillary flow measurements sug-gested that a clinical verification of theanalogue studies by Wasserman et al. (I966)would be worth while. For in patients withvarious forms of heart disease leading topulmonary hypertension, the capillary flowpulsatility index became attenuated when thepulmonary arterial resistance rose (Karatzasand Lee, I966). We naturally hoped thatsimply measuring the pulsatility index byrecording N20 uptake in the body plethysmo-

3

graph would prove to be a useful noninvasivemethod for deriving pulmonary arterialresistance. Moreover, the shape of the normalcapillary flow curve was so similar to the pul-monary artery- left atrial pressure difference,responsible for propelling blood through thelung capillary system, that we also hoped foruseful diagnostic aid by comparing the N20uptake curve with the PA- pulmonary wedgepressure differences, in order to obtain similarnoninvasive evidence of instantaneous pres-sure events in these two sites by recordingtheir effect on capillary blood flow. This wehoped would be particularly useful in differen-tiating stenosis from incompetence in patientswith mixed mitral valve disease. Our hopeswere too sanguine, as is shown by our resultsfrom 36 patients with heart disease leading tovarious combinations of pulmonary arterialand pulmonary venous hypertension (Karatzasand Lee, I970). Fig. 3 shows the results weobtained. It can be seen that there was anegative correlation between the pulmonarycapillary pulsatility index (Qc max/Qc) andeach of the following: (i) mean pulmonaryarterial pressure; (2) left atrial pressure,measured indirectly as the pulmonary arterialwedge pressure; and (3) the pulmonaryarterial resistance.These findings showed that patients with

heart disease affecting the pulmonary circula-tion developed abnormal pulmonary capillaryflow patterns, but that the degree and type oflung capillary flow abnormality could not beused alone to predict either the severity or theform of the pulmonary vascular diseasepresent. For, though patients with raisedpulmonary vascular resistance tended to havelow pulsatility indices, so did patients withpoor right ventricular ejection, low strokevolume, or tachycardia. Moreover, patientswith high left atrial pressure but withoutsevere pulmonary arterial hypertension alsotended to have capillary flow pulses of dimin-ished amplitude with an abnormal configura-tion, particularly those patients with pre-dominant mitral incompetence. Thus anincrease in pulmonary arterial resistance or anincreased pulmonary venous input impedance,or a combination of the two, could produceattenuation of the pulmonary capillary flowpulse. Subsequently we have found a fewexamples of patients with high pulmonaryarterial pressures in whom the capillary flowpulsatility index has remained remarkablynormal. This reminded us of findings byFishman and his colleagues (Morkin, Levine,and Fishman, I964), who had studied theeffects of induced pulmonary hypertension byhypoxia on pulmonary capillary flow pulsa-



j.com/

Br H


ownloaded from


20 G. de Y. Lee

- 35-I

EE 30

fi 25

Q0.20

a,-0~034

15

0 10

C40 -

0~

10 I-2 I-4 I6 1-8 2-0 2-2 2-4°c max./z5c

0

1-0 1-2 1-4 1: 1-8 2:0 2-2 24Qc max./Oc

1-0 1-2 1-4 1-6 1-8 2-0 2-2 2-4Qc max./°c

FIG. 3 Correlation between the ratio of peak to mean capillary flow rates and (a) meanpulmonary artery pressure, (b) mean pulmonary arterial wedge pressure, and (c) pulmonaryarterial resistance in 36 patients with heart disease.

tility in the dog. They found that in spite ofanobvious rise in pulmonary vascular resistancethe pulsatility index remained unaltered withhypoxia.We thought that the explanation of both

this finding and our clinical finding that thepulsatility index sometimes remained normalat high pulmonary artery pressures could bedue to a combined effect of a simultaneous re-duction of pulmonary arterial distensibility orcompliance (C) occurring in conjunction withthe increased resistance (R) responsible forthe elevated pressures in both instances. If (C)were to fall as (R) were rising, then the timeconstant of the system (C x R) would remainconstant over a wide range of pulmonaryarterial pressure, thus preserving an un-

altered flow pulsatility in the capillaries.We therefore tested this hypothesis in

animals. The pulmonary capillary blood flowwas measured in the body plethysmographsimultaneously with pulmonary arterial inflowusing a cuff electromagnetic flowmeter, pul-monary artery pressure, left atrial pressure,and computed pulmonary arteriolar pressurein closed-chested atrially paced dogs. The leftatrial pressure was below 5 cm. of H20 in allinstances. The effect of tachycardia, hypoxia(IO%0 02), and serotonin infusion was studied.The pulmonary arterial volume was calculatedfrom the ether circulation time from aninjection site at the pulmonary valve to thelung capillaries in the body plethysmograph(Feisal, Soni, and DuBois, I962; Sackner,Will, and DuBois, I966). We next calculated

the pulmonary arterial resistance convention-ally and obtained pulmonary arterial com-pliance (C) by dividing the systolic storagevolume (PA inflow- Pc outflow during sys-tole = AV) by the mean PA systolic distendingpressure ((PAp + P arteriolar p)/2 = AP).The pulmonary arterial distensibility (D)

was calculated as the ratio of the lumpedcompliance (C) to the mean pulmonaryarterial volume (Vo) as determined by theether plethysmograph method and expressedas a percentage volume change per unitpressure thus:

D= I00D=AP x Vo xIO (i)

where AV= systolic pulmonary arterialstorage volume (ml.)

AP = pulse pressure oflumped pul-monary arterial distendingpressure (cm. H20)

Vo = mean pulmonary arterial vol-ume (ml.).

The pulmonary arterial pulse wave velocitycould also be calculated from the Bramwell-Hill equation:

Co=Vo

(ii)

where Co = PA pulse wavevelocity(cm./sec.)Vo =PA volume (ml.: from stroke

volume and ether time)C =PA compliance (ml./cm. H20)

but from equation (i)V ioo

C=

D

70

M 60-EE* 50-L,

L- 40-a.

v 30

L. 20-0a03 10-

0

0

0

..

*

.

N\



j.com/

Br H


ownloaded from


Regulation ofpulmonary circulation 21

where D = lumped PA distensibility (%vol. change/cm. H20)

so Co=AID(

Fourier analysis of pressure and flow datawas also undertaken in order to examine thepulmonary arterial input impedance changesimposed by hypoxia and serotonin.

Fig. 4 shows the results of our compliancestudies in I3 dogs. We found that there was ahyperbolic relation between the rising pul-monary arterial resistance and the fallingpulmonary arterial compliance induced bygraduated changes in hypoxia and serotonininfusion. Fig. 5 next shows that the pulmon-ary arterial time constant (t) remained uniformover a very wide range of pulmonary arterialpressure induced by either hypoxia or sero-tonin infusion. Only above a mean pulmonaryartery pressure of 40 mm. Hg did the PA timeconstant begin to increase. The time constant(t) was calculated from the product ofpulmon-ary arterial compliance (C) and pulmonaryarterial resistance (R). This finding gave asatisfactory explanation for the maintainedpulsatility of lung capillary blood flow in con-ditions of moderate pulmonary arterialhypertension, for as precapillary resistance in-creased, tending to attenuate lung capillaryflow pulsatility, its effect was counteracted bythe decreased compliance of the pulmonaryarterial system, tending to increase trans-mission of pulsatility to the capillaries. Theeffects of the decreasing distensibility of thepulmonary arterial system, produced by in-creasing degrees of pulmonary hypertension,had a natural consequence of increasing thepulmonary arterial pulse wave velocity. Therelationship between the mean pulmonaryarterial pressure and pulse wave velocity isshown in Fig. 6. Full details of this studyhave been published (Reuben et al., 197oa).The pulmonary arterial input impedance

calculations which we obtained from ourinstantaneous records of pulmonary arterialpressure and pulmonary capillary blood flowyielded fresh information about the site ofresistance to blood flow in the arterial systemof the lungs. We also found that the pulmon-ary arterial input impedance spectrum wasindependent of heart rate, which implies thatthe pulmonary arterial system behaves moder-ately linearly, for in a linear system an altera-tion in the harmonics of pressure by increasedheart rate would be accompanied by identicalchanges in the harmonics of flow so that theratio of the two, namely the impedancemodulus, would still lie on the control slopebut at a new point appropriate to the higher

3-2

30

2.8

2.6

2-4

E 22 -r-.J

E1-8- 8

I6)1-4E

'u 1-6

a-

0-4

0'2

0--

0.8

oPA

0 0

0o control

x = hypoxiao = ser-otonin

S.

Ox x

0X X

XX00X

Ib a I

66 0-7 0-8 0-9 1.0

* .0

0X eOX X X

O.0 0~~%6i*0

0- -o 0

041 02 0.3 0.4 Q05resistance (cm. H20/mi. per sec.)

FIG. 4 The relationship between pulmonaryarterial resistance and the corresponding valueof compliance in dogs. The graph combines datafrom control (*), hypoxia (x ), and serotonin(0 ) studies.

fundamental frequency imposed by thechange in heart rate.The pulmonary arterial input impedance

spectrum during air breathing showed a firstminimum value of ioo dyn.sec.cm. 5 occur-

ring at 3-52 cycles per second (c.p.s.). A firstmaximum of 2I4 dyn.sec.cm. took place at

9go c.p.s., followed by a second minimum at

I3'0 c.p.s. The impedance phase angles were

negative at low frequencies, indicating flowleading pressure, but they became positiveabove 3-0 c.p.s. The reflection coefficientduring this period was calculated to be 03Iand the average distance from the pulmonaryvalve to the reflecting site was I3-5 cm.

Hypoxic ventilation caused a rise in the in-put resistance from 540 to 68o dyn.sec.cm. -I

and a shift in the frequency of the first im-pedance minimum to 5 6o c.p.s. The reflectioncoefficient for these data was calculated to beo049, while the distance to the reflecting site

x

x8;000o 40o

X0Xx



j.com/

Br H


ownloaded from


22 G. deJI. Lee

j1

@0e7-~~~~~~~~~~~° °4- 04^

3

tcompli

E~~~~~~~~~~~~~~~~ 1

6~~~

2E

41~ t4

UC

0-.I0.~~~~E i0u

p

0 -50 15PAp cm.H20

iance

30 45

FIG. 5 The relationship between pulmonaryarterial compliance (C), pulmonary arterialresistance (R), mean pulmonary artery pres-sure, and pulmonary arterial time constant (t)obtainedfrom the product of (C x R) in dogs.Data from control hypoxia and serotoninstudies were combined to obtain these relation-ships.

fell to I0o2 cm. Serotonin infusion caused aneven greater rise in the input resistance to85o dyn.sec.cm.-5 and a bigger shift of thefirst impedance minimum to 6'6o c.p.s. Thereflection coefficient was calculated to be o053and the distance to the reflecting site to beI2'2 cm.

The most dramatic change in the pulmon-ary arterial input impedance spectrum wasproduced by sympathetic nerve stimulationundertaken in two dogs. The input impedancerose to 780 dyn.sec.cm. - 5 and the firstimpedance minimum shifted to 7.5 c.p.s.,with a consequent large fall in the distance tothe reflecting site to 8-7 cm. and a rise in thereflection coefficient to os6. The results ofthese studies have been published (Reubenet al., 197ob).These results indicate that acute pulmonary

hypertension causes marked alterations in thedistal pulmonary arterial bed characterized byan increased vascular resistance. The increasein wave reflection, and also the shift proximallyof the site of wave reflection towards the pul-monary valve, indicate that the main site ofresistance to blood flow in the pulmonary

FIG. 6 The relationship between pulmonaryarterial pulse wave velocity and mean pulmon-ary arterial input pressure (PAp). The pulsewave velocity was calculatedfrom the com-pliance data shown in Fig. 4 using theBramwell-Hill equation.

arterial system is capable of moving severalcm. proximally into the larger vessels of thepulmonary arterial system. This implies thathypoxia, serotonin, and sympathetic nerveaction all have effects on the larger vessels ofthe pulmonary arterial system as well as at themore conventional resistance site in thearterioles.

Previous studies by Ingram et al. (I968) onisolated lobes of the lung in situ, with theirnerve supply intact, had already shown thatduring sympathetic stimulation the largerpulmonary arteries became stiffer, whereasthe calculated pulmonary arterial resistanceincreased only slightly. In contrast nor-adrenaline infusion produced less stiffnesschange in the larger pulmonary arteries butgreater effects on the arterioles. Furtherstudies (Ingram, Szidon, and Fishman, I970)have shown that hypothalamic stimulation,acting via the sympathetic outflow, showed apredominant action upon the media of thelarger pulmonary arteries, where the sympa-thetic nerve endings are located. These find-ings by Fishman's group have direct relevanceto our pulmonary artery compliance studiesfor they show the mechanisms whereby thepulmonary arterial compliance is regulated in

4

3

_-: 3EU

x

u72-g2l

0

41:0.

_0

-4

2

-o3

m1,

a,A.-

Q L0iA-p cm.H20

20LO



j.com/

Br H


ownloaded from



by cardiac catheterization simultaneously withN20 uptake in the body plethysmograph. Hethen correlated the pulmonary artery flowconduction time (Ct) against mean pulmonaryartery pressure. He found a curvilinear re-lationship between the two and was able toderive the mean pulmonary arterial pressurewith great accuracy from measurements of theflow conduction time using nitrous oxide(Reuben, I970). Reuben showed that therelationship between flow conduction timeand mean pulmonary artery pressure con-formed to the following equation:

Ct=3Io +2910PAp-2-2

0

0 20Mean PA pressure (mm.Hg)

FIG. 7 The relationship betwearterial conduction time (Ct) an

pulmonary artery pressure (PAl

pulmonary hypertension andcreasing compliance is not a

effect due to vascular distensicThe present consensus of opi

the actions of chemical ageniupon the pulmonary circulsuggests that hypoxia, pressora lesser extent hypercapnoealocal vasocontrictor effects upoand venules of the lung, but inresponses are excited which inoness of the more proximal pulnThe problem now facing invdesign studies which will allcimportance of reflex and lochassessed in their relationshipwithin the pulmonary circulatiThe increased pulse wave ve

monary artery flow conducti(were found by Reuben et ai

which were associated with,monary arterial compliance asartery pressure rose in animalReuben to explore the possibilconduction time as a meansmean pulmonary artery presN20 uptake curve in patientplethysmograph. He thereforKaratzas and Lee's data (I969)patients with heart disease inmonary arterial pressure had I

Role of the pulmonary venous system inregulating blood flow in the lung capil-laries

40 60 Having briefly reviewed the regulation of the

dynamics of blood flow into the lung capillary

renpulmonary system by the pulmonary arteries, and how the

ed mean alveolar capillary systems themselves contri-

p) in man. bute to regulating pulmonary vascular pres-

sure and flowdistibution in the lungs, it nowremains necessary to examine the role of thepulmonary venous system.

that the de- Under normal conditions the N20 uptakepurely passive curve, representing pulmonary capillary bloodn. flow, virtually resembles the profile of the pul-inion regarding monary arterial pressure curve though it ists and reflexes separated from it in time. Blood flow throughlation strongly the lung capillaries is therefore largely un-

amines, and to influenced by pressure events occurring in theproduce direct venous system of the lungs, which implies that)n the arterioles pressure events from the left atrium are

addition reflex damped out from reaching the capillaries be-crease the stiff- cause of the anatomical characteristics of thenonary arteries. pulnonary venous system. Part of theestigators is to explanation for this may be inferred fromcw the relative work on the isolated rabbit lung by Caro andal action to be Saffnan (I965). These workers showed thatto one another the pulmonary venous system has an ex-

ion. tremely non-linear compliance, being highly.locity and pul- distensible at low venous pressures buton timne which rapidly becoming indistensible as the venous

1. (I97oa), and pressure rises. Harris, Heath, and Apostolo-decreased pul- poulos (I965) had examined the distensibilitythe pulmonary of the pulmonary arterial trunk in man and[s, have excited had found that it was considerably more dis-ity of using the tensible than the aorta. They used circum-of calculating ferential strips of these vessels, obtained at

-sure from the necropsy, and plotted length tension curves

:s in the body from them, using a travelling microscope aftere re-examined setting up the material in a tension balance.obtained from They did not extend their work to the pulmon-whom the pul- ary veins. I have carried out limited studies,been measured using their technique on the pulmonary

200

-1~

E1500"IO

Ec0

C0uo

0

0~

* 50-< s

0

0

.



j.com/

Br H


ownloaded from


24 G. deJ. Lee

venous trunk in necropsy material and havecompared it with the pulmonary arterial find-ings of Harris. My findings were identical withthose ofHarris and his associates with regard tothe pulmonary artery. However, the pulmon-ary vein, though initially highly distensible,became virtually indistensible when its tensionexceeded 2-5 dynes/mm.2 In terms of pressurerelated to vessel dimension, this means that atpressures exceeding IO-I5 mm. Hg a state ofaffairs will be reached where the pulmonaryveins attain the limits of their distensibility.Pressure events occurring in the left atriumare then likely to be transferred retrograde tothe venular end of the capillary system, havingbeen damped out previously, and might beexpected to impede lung capillary outflow ina manner which could be predicted from theshape of the left atrial pressure curve, pro-vided that the arterial inflow profile wasnormal. In order to test the hypothesis thatepisodic changes in pulmonary venous im-pedance during the cardiac cycle might affectthe pulsatility of lung capillary blood flow Ihave found it convenient to examine a some-what contrived situation where these eventsare taking place in virtual isolation. Thishappens in patients with complete heartblock in whom the ventricular rate is bothfixed and slow and atrial contraction takesplace at a faster rate, which is also randomlyrelated to ventricular events (Gillespie et al.,I967). In this way episodic changes in pulmon-ary venous impedance were produced by thecontractions of the left atrium, which variedrandomly throughout the ventricular cycle.During ventricular diastole the effects ofpulmonary venous pressure fluctuation trans-mitted from the left atrium could be studiedby searching for alterations in the normal lungcapillary blood flow profile in relation to theP wave of the electrocardiogram. In patientswith congenital complete heart block, wherethe left atrial pressure was normal, the leftatrial pressure events produced very smallchange in lung capillary blood flow. However,when the left atrial pressure was high, as inacquired complete block with heart failure, anentirely different capillary blood flow profilewas obtained. Under these circumstances, withhigh left atrial pressures, cannon waves fromthe left atrium were transmitted retrogradethrough the pulmonary venous system to thecapillaries and at times could be picked up bypressure monitoring in the pulmonary artery.Under these circumstances N20 uptake fromthe lung capillaries virtually ceased at thetimes when episodic increases in venous out-flow impedance were produced by the cannonwave (Lee, I969a).

More recently I have been studying theeffect of isolated mitral incompetence uponlung capillary blood flow pulsatility. Inpatients where the main pulmonary arterypressure is normal at rest but who have severemitral incompetence resulting in giant leftatrial V waves, the left atrial V wave mayapproach some 40 mm. Hg and be conductedretrograde to the pulmonary arterial system.During these periods N20 uptake from thelungs virtually ceases, and indeed lungcapillary blood flow in mitral incompetenceunder these circumstances virtually dependson a long diastolic interval for resumption ofblood flow from the pulmonary artery to thepulmonary veins (Lee I969b). Normal pulsa-tile lung capillary blood flow events may bereinstated by replacement of the leakingmitral valve by means of a Starr-Edwards ballvalve. These N20 studies provide an explana-tion for surprising dyspnoea sometimes en-counteredinpatients withpure mitralincompe-tence and normal pulmonary arterial pressuresat rest, who become intensely breathless onexertion if this is associated with a rapid in-crease in heart rate. Heart rate control withdigitalis and propranolol is often dramaticallybeneficial in such cases, because of the in-creased diastolic flow time through the lungcapillaries produced by good heart rate control.These clinical observations allow only intui-

tive deductions to be made about the role ofthe pulmonary veins and left atrium inmodifying the transmission of pressure andflow through the lung vasculature. Precisephysiological study of the role of the pulmon-ary veins in regulating lung capillary flow arevirtually non-existent at the time of preparingthis review. However, Caro, Bergel, and Seed(I967) have examined the transmission ofpressure waves through the pulmonary vascu-lar bed. They measured pressure in the lobarpulmonary arteries, veins, and left atria ofanaesthetized open-chested dogs. Observa-tions were made before and after snaring thelobar vessels and before and after infusions ofdextran to raise the mean pressure. Forwardand backward transmission (output/inputpressure) and transmission ratios (forward/backward transmission) were calculated fromthe pressure waves submitted to Fourieranalysis. At low frequencies the vascular bedwas non-symmetrical, forward transmission(arteries to veins) being several times greaterthan backward transmission (veins to arteries)at the same pressure. These findings wouldsuggest that at low pulmonary vein pressurethe PA pressure and flow pulse will bedamped out in its passage through the venoussystem so that it becomes more uniform. We



j.com/

Br H


ownloaded from



have been able to confirm this in our labora-tory, measuring pulmonary vein inflow to theleft atrium using an electromagnetic flow-meter. Our findings confirm the findings ofKinnen and Stankus (I968), who showed thatthe velocity of blood flow in the pulmonaryveins entering the left atrium is largely uniformbut also influenced by the pressure eventstaking place locally in the left atrium. How-ever, the work of Caro et al. (I967) also sug-gests that if the left atrium and pulmonaryvenous pressure was high, so that the veinsbecame non-compliant, then arterial pulsa-tions of pressure and flow would be trans-mitted across the capillary system into theveins so that a more 'arterial' venous flowpulse would be observed. It is interesting thatMorkin et al. (I965) had described suchpulsatility of venous flow by direct measure-ments in dogs and had concluded that thisvenous pulsatility was transmitted from theright heart through the arterial and capillarysystems to dominate venous flow. It wasrelevant that their measurements were con-ducted at moderately high pulmonary venouspressures (io mm. Hg). Controversy soonarose about the discrepancies in flow velocityprofile found by various workers in the pul-monary veins. Our current studies in animalsindicate that the mean level of pressure in theleft atrium and pulmonary veins dominatesthe behaviour of venous outflow impedance toblood flow from the lung capillaries. Thus,when left atrial pressure is low, pulmonaryvenous inflow to the left atrium is moderatelyuniform, being momentarily impeded by leftatrial systole. However, if the left atrial pres-sure is elevated by inflating a water-filledballoon in the left atrium of closed-chesteddogs the pulmonary venous flow to the leftatrium becomes highly pulsatile and resemblesa damped lung capillary blood flow pulsation,delayed in time from capillary events. Con-siderable further work is needed before wewill have acquired as precise a knowledge ofvenous influences upon lung capillary bloodflow as we now possess with regard to arterialfunction in the lungs.

ReferencesBanister, J., and Torrance, R. W. (I960). The effects

of the tracheal pressure upon flow: pressure rela-tions in the vascular bed of isolated lungs. QuarterlyJournal of Experimental Physiology and CognateMedical Sciences, 45, 352.

Bates, D. V., Varvis, C. J., Donevan, R. E., andChristie, R. V. (I96o). Variations in the pulmonarycapillary blood volume and membrane diffusioncomponent in health and disease. Journal of ClinicalInvestigation, 39, I40I.

Bosman, A. R., Honour, A. J., Lee, G. de J., Marshall,R., and Stott, F. D. (I964). A method for measuringinstantaneous pulmonary capillary bloodflow and

right ventricular stroke volume in man. ClinicalScience, 26, 247.

Caro, C. G., Bergel, D. H., and Seed, W. A. (I967).Forward and backward transmission of pressurewaves in the pulmonary vascular bed of the dog.Circulation Research, 20, I85.

-, and Saffman, P. S. (I965). Extensibility of bloodvessels in isolated rabbit lungs.Journal ofPhysiology,178, I93.

Cudkowicz, L. (I968). The Human Bronchial Circula-tion in Health and Disease. Williams and Wilkins,Baltimore.

Feisal, K. A., Soni, J., and DuBois, A. B. (I962).Pulmonary arterial circulation time, pulmonaryarterial blood volume, and the ratio of gas to tissuevolume in the lungs of dogs. J'ournal of ClinicalInvestigation, 41, 390.

Gillespie, W. J., Greene, D. G., Karatzas, N. B., andLee, G. de J. (I967). Effect of atrial systole on rightventricular stroke output in complete heart block.British Medical3Journal, I, 75.

Glazier, J. B., Hughes, J. M. B., Maloney, J. E., Pain,M. C. F., and Webb, J. B. (I966). Decreasingalveolar size from apex to base in the upright lung.Lancet, 2, 203.

, , -, Rosenzweig, D. Y., and West, J. B.(I968). Effect of vascular pressures on pulmonarycapillary morphology. Clinical Research, I6, 370.

Harris, P., Heath, D., and Apostolopoulos, A. (I965).Extensibility of the human pulmonary trunk.British HeartJournal, 27, 65I.

Heinemann, H. O., and Fishman, A. P. (I969). Nonrespiratory functions of mammalian lung. Physio-logical Reviews, 49, I.

Howell, J. B. L., Permutt, S., Proctor, D. F., andRiley, R. L. (I96I). Effect of inflation of the lungon different parts of the pulmonary vascular bed.Journal of Applied Physiology, I6, 71.

Hughes, J. M. B., Glazier, J. B., Maloney, J. E., andWest, J. B. (I968). Effect of lung volume on thedistribution of pulmonary blood flow in man.Respiration Physiology, 4, 58.

Ingram, R. H., Jr., Szidon, J. P., and Fishman, A. P.(1970). Response of the main pulmonary artery ofdogs to neuronally released versus blood-bornenorepinephrine. Circulation Research, 26, 249.

,- , Skalak, R., and Fishman, A. P. (I968).Effects of sympathetic nerve stimulation on thepulmonary arterial tree of the isolated lobe perfusedin situ. Circulation Research, 22, 8oi.

Karatzas, N. B., and Lee, G. de J. (I966). The effect ofpulsatile capillary blood flow on gas exchangewithin the lungs; a consideration of some hydro-dynamic factors which affect this pulsatility in man.Bulletin de Physio-pathologie Respiratoire, 2, 521.-, (I969). Propagation of blood flow pulse in

the normal human pulmonary arterial system.Analysis of the pulsatile capillary flow. CirculationResearch, 25, II.

- (1970). Instantaneous lung capillary bloodflow in patients with heart disease. CardiovascularResearch, 4, 265.

- , - , and Stott, F. D. (I967). A new electro-pneumatic flowmeter for the body plethysmograph.Journal of Applied Physiology, 23, 276.

Kinnen, E., and Stankus, A. J. (I968). Pulmonaryvenous blood flow. U.S.A.F. Report No. SAM-TR-68-22.

Lee, G. de J. (I969a). Blood flow in the lungs. InModern Trends in Cardiology, vol. 2, p. 197. Ed. byA. M. Jones. Butterworths, London.

- (I969b). Lung blood flow studies in man usingthe whole body-plethysmograph: technical aspectsand applications. Progress in Respiration Research,4, I40.



j.com/

Br H


ownloaded from


26 G. deJ. Lee

Lee, G. de J., and DuBois, A. B. (I955). Pulmonarycapillary blood flow in man. Journal of ClinicalInvestigation, 34, 1380.

-, Matthews, M. B., and Sharpey-Schafer, E. P.(1954). The effect of the Valsalva manoeuvre onthe systemic and pulmonary arterial pressure inman. British Heart Journal, I6, 3II.

Mead, J., Takishima, T., and Leith, D. (I967).Mechanical interdependance of distensible units inthe lung. Federation Proceedings, 26, 55I.

Menkes, H. A., Sera, K., Rogers, R. M., Hyde, R. W.,Forster, R. E., II, and DuBois, A. B. (I970).Pulsatile uptake of CO in the human lung. Journalof Clinical Investigation, 49, 335.

Milic-Emili, J., Henderson, J. A. M., Dolovich, M. B.,Trop, D., and Kaneko, K. (I966). Regional dis-tribution of inspired gas in the lung. Journal ofApplied Physiology, 21, 749.

Morkin, E., Collins, J. A., Goldman, H. S., andFishman, A. P. (I965). Pattern of blood flow in thepulmonary veins of the dog. Journal of AppliedPhysiology, 20, I I I 8.

-, Levine, 0. R., and Fishman, A. P. (I964).Pulmonary capillary flow pulse and the site ofpulmonary vasocontriction in the dog. CirculationResearch, I5, I46.

Permutt, S. (i965). Effect of interstitial pressure of thelung on pulmonary circulation. Medicina Thoracalis,22, II8.

-, Bromberger-Barmea, B., and Bane, H. N. (I962).Alveolar pressure, pulmonary venous pressure, andthe vascular waterfall. Medicina Thoracalis, 19, 239.

Reuben, S. R. (I970). Wave transmission in thepulmonary arterial system in disease in man.Circulation Research, 27, 523.

-, Gersh, B. J., Swadling, J. P., and Lee, G. de J.(I970a). Measurement of pulmonary arterial dis-tensibility in the dog. Cardiovascular Research 4,473.

- , Swadling, J. P., Gersh, B. J., and Lee, G. de J.(197ob). Impedance and transmission properties ofthe pulmonary arterial system. CardiovascularResearch, 5, i.

Rosenzweig, D. Y., Hughes, J. M. B., and Glazier, J. B.(1970). Effects of transpulmonary and vascularpressures on pulmonary blood volume in isolatedlung. Journal of Applied Physiology, 28, 553.

Roughton, F. J. W., and Forster, R. E. (I957). Relativeimportance of diffusion and chemical reaction ratesin determining rate of exchange of gases in thehuman lung, with special reference to true diffusingcapacity of pulmonary membrane and volume ofblood in the lung capillaries. Journal of AppliedPhysiology, II, 290.

Sackner, M. A., Will, D. H., and DuBois, A. B. (I966).The site of pulmonary vasomotor activity duringhypoxia or serotonin infusion. Journal of ClinicalInvestigation, 45, I12.

Wasserman, K., Butler, J., and Van Kessel, A. (I966).Factors affecting the pulmonary capillary flowpulse in man.Journal ofApplied Physiology, 2I, 890.

Weibel, E. R. (I963). Morphometry of the Human Lung.Springer, Berlin.

West, J. B., and Dollery, C. T. (I960). Distribution ofblood flow and ventilation-perfusion ratio in thelung, measured with radioactive carbon dioxide.journal of Applied Physiology, IS, 405.

, , and Naimark, A. (I964). Distribution ofblood flow in isolated lung: relation to vascular andalveolar pressures. Journal of Applied Physiology,19, 7I3.



j.com/

Br H


ownloaded from


journal, regulation ofthe pulmonary circulation

Documents