left ventricular myocardial edema - circulation...

1713

Left Ventricular Myocardial EdemaLymph Flow, Interstitial Fibrosis, and Cardiac Function

Glen A. Laine and Steven J. Allen

We hypothesized that both acute and chronic accumulation of myocardial interstitial edema(extravascular fluid [EVF]) would compromise cardiac function. We also postulated that excess

fluid within the myocardial interstitial space would potentiate interstitial fibrosis, thus furthercompromising function. Dogs were divided into three groups: 1) control, 2) chronic pulmonaryhypertensive with right heart failure, and 3) chronic arterial hypertensive. The quantity of EVF,expressed as the unitless blood-free (wet weight-dry weight)/dry weight ratio, and interstitialfibrosis (collagen content) were determined and correlated with cardiac function at baselineand after acute elevation of coronary venous pressure and reduction of cardiac lymph flow.Control EVF was 2.90±0.20 (mean±SD), which increased to 3.45±0.16 after acute (3-hour)elevation of coronary sinus pressure. This EVF significantly compromised cardiac function. TheEVF in chronically hypertensive dogs and in dogs with chronic right heart pressure elevationswas 3.50+0.30 and 3.50±0.08, respectively. End-diastolic left ventricular interstitial fluidpressure increased from a control value of 14.9+3.1 (at EVF=2.9) to 24.8+3.7 (at EVF=3.5).An EVF of 3.5 produced -30% reduction of the heart's ability to maintain cardiac output at a

left atrial pressure of 15 mm Hg. The compromised function in these chronic models isexacerbated after acute elevation of coronary venous pressure and reduction of cardiac lymphflow. Collagen levels were elevated by at least 20% in the chronic hypertensive dogs and in thenonhypertrophied left ventricles of dogs with chronic right heart pressure elevation. Weconclude that acute myocardial edema compromises cardiac function and that chronic rightheart pressure elevation and chronic arterial hypertension produce left ventricular myocardialedema, which also compromises function in these common pathological conditions. Thepresence of myocardial edema in these chronic models also potentiates interstitial fibrosis,leading to a further decrease in the heart's ability to function normally. (Circulation Research1991;68:1713-1721)

T he presence of excess myocardial interstitialfluid or myocardial edema increases the stiff-ness and decreases the compliance of the left

ventricle.1-3 The accumulation of myocardial edemahas been demonstrated in a number of acute andchronic conditions, including cardiac transplanta-tion,4 decreased plasma colloid osmotic pressureduring cardioplegia,5 altered microvascular perme-ability by chronic arterial hypertension,6 elevation ofmyocardial microvascular pressures,7 and chronicreductions in myocardial lymph flow (QLV) rate.8 Thepresence of chronic edema within the interstitium

From the Center for Microvascular and Lymphatic Studies andthe Department of Anesthesiology, The University of Texas Med-ical School, Houston, Tex.

Supported by National Heart, Lung, and Blood Institute grantsHL-36115 and HL-01999. G.A.L. is an Established Investigator ofthe American Heart Association.Address for correspondence: Dr. Glen A. Laine, Department of

Anesthesiology, The University of Texas Medical School, 6431Fannin, MSMB 5.020, Houston, TX 77030.

Received November 6, 1989; accepted March 5, 1991.

has also been shown to stimulate fibrosis within themyocardial interstitial matrix.9-1' The increased dep-osition of collagen within the myocardial interstitialspace during chronic arterial hypertension and ven-tricular hypertrophy compromises cardiac function.12We believe this deposition of collagen may resultfrom the combined effects of cardiac interstitialremodeling during hypertrophy and the presence ofsmall volumes of myocardial edema.6'13We hypothesized that acute and chronic accumu-

lation of myocardial edema could directly compro-mise cardiac function in a predictable manner. Wealso postulated that myocardial fibrosis secondary tosmall volumes of chronic edema could diminish theheart's ability to function normally. Both the pres-ence of myocardial edema and the edema-induceddeposition of interstitial matrix material could in-crease myocardial stiffness, increase oxygen diffusiondistances, and alter the structural elements of theinterstitium, thus compromising cardiac function un-der both control and stressed conditions.

by guest on May 15, 2018

http://circres.ahajournals.org/D

ownloaded from

http://circres.ahajournals.org/

1714 Circulation Research Vol 68, No 6 June 1991

Materials and MethodsAnimal Preparation

Forty-eight mongrel dogs (drawn from acute andchronic populations) with body mass >15 kg and ofeither sex were used. Anesthesia was induced withthiopental sodium (20 mg/kg) and maintained with0.5-1.5% halothane. The dogs were intubated andartificially ventilated with a respirator (Harvard Ap-paratus, South Natick, Mass.) set to deliver room airat a volume of 25 ml/kg and at a rate appropriate tomaintain Paco2 between 35 and 40 mm Hg. Fluid-filled catheters were placed into the femoral artery,femoral vein, and left jugular vein. A Fogarty bal-loon-tipped catheter (8/22F, 43.0 ml), cardiac outputthermodilution catheter (7F), and Swan-Ganz (5F)catheter (all from Edwards Laboratories, Santa Ana,Calif.) were placed into the right jugular vein. Thechest was then opened using a midline incision, andthe major lymphatic trunk draining the left ventricu-lar myocardium was identified and cannulated aspreviously described.'415 With the chest open, theFogarty balloon was secured in place just above theentry of the azygos vein into the superior vena cava.The thermodilution catheter was advanced into thepulmonary artery, and the Swan-Ganz balloon-tipped catheter was sutured into the coronary sinusin a nonocclusive manner as described previously.15A solid-state microtipped pressure transducer(Millar, Houston) was placed into the left carotidartery and advanced into the left ventricular cavity. Afluid-filled catheter was also inserted into the leftatrial appendage and secured with a purse suture.

Five dogs from this group had chronic (2-month)elevation of right ventricular and central venouspressures. We produce chronic right ventricular pres-sure elevations by banding the pulmonary artery.16-18As right heart pressures increase, pressure within thethebesian veins and coronary sinus increases. Thiselevates microvascular pressure within the left ven-tricle, thus potentiating left ventricular myocardialedema. As microvascular fluid and protein flux in-crease, a phenomenon referred to as "wash-down"19,20 reduces the concentration of plasma pro-teins in the interstitial fluids, resulting in low-proteinedema. Elevation of superior vena caval pressure(SVCP) in this model also reduces the volume ofcardiac lymph that can flow into the central venouscirculation.21 This model produces right-sided pres-sures within the heart and vasculature analogous tothose seen in right heart failure or cor pulmonale.

Seven dogs were drawn from a chronic populationof one-kidney, one-clip Goldblatt hypertensive dogsand prepared in the preceding manner. These dogshad a sustained elevation of systemic arterial pres-sure for 9±1 weeks before their use in the acutepreparation.19 These hypertensive dogs have beenshown to exhibit a significant increase in microvascu-lar permeability over time.6,7 The increase in perme-ability allows plasma proteins from the microcircula-tion to enter the interstitial space more easily and to

produce relatively high-protein edema. In 18 of thedogs from the various groups, porous polyethylenecapsules were chronically implanted in the left ven-tricular myocardium for the measurement of end-diastolic interstitial fluid pressure.6.'5 Three porouspolyethylene capsules (35 -gm pores) were implantedin each left ventricular myocardium. Capsules wereplaced in the midmyocardium, since interstitial hy-draulic continuity produces rapid hydraulic equilib-rium within the interstitial space." When viewedunder the microscope, sections of tissue that con-tained capsules revealed normal myocytes surround-ing loose connective tissue, which had encased theporous polyethylene.

Physiological Measurements and Tissue AnalysisAll fluid-filled catheters were connected to

Statham pressure transducers (model P23Db, HatoRey, Puerto Rico), and all data were recorded on achart recorder (Grass Instrument Co, Quincy,Mass.). The femoral artery catheter was used tomeasure systemic arterial pressure, and the femoralvein catheter was used to administer fluids or drugsas necessary. The catheter within the left jugular veinwas used to measure SVCP above the azygos vein andFogarty balloon. The Swan-Ganz balloon catheterwithin the coronary sinus was used to record andmanipulate coronary sinus pressure (CSP). Lymphfrom the left ventricle was allowed to flow through acalibrated pipette and was returned to the superiorvena caval system through a set of pressure-tightfittings. This allowed the evaluation of true QLV asSVCP varied.22 This arrangement emulates the nor-mal in vivo circumstances in which myocardial lymphdrains into the superior vena caval system. Lymphsamples were obtained by temporarily disconnectingthe pressure-tight lymphatic catheter and allowinglymph to flow into heparinized test tubes. Cardiacoutput was repetitively determined by injection of 3ml cold Ringer's solution through the thermodilutioncatheter, and the data were recorded from an Ed-wards 9520 cardiac output computer. Cardiac func-tion curves were generated by recording cardiacoutput and left atrial pressure (from the previouslyimplanted left atrial pressure catheter) as incremen-tal volumes of Ringer's and albumin solutions wereinfused. Albumin was added to maintain plasmacolloid osmotic pressure near control levels. Thisprevented the accumulation of interstitial fluid re-sulting from a decrease in colloid osmotic pressureduring cardiac output determinations.The amount of myocardial edema or extravascular

fluid (EVF) was obtained from the unitless blood-free (wet weight- dry weight)/dry weight ratio for theleft ventricle. After removal of tissue for biochemicaland histological analysis, the myocardium wasweighed and homogenized, and a lipid extractionwith a 2:1 mixture of chloroform and methanol wasperformed. The homogenate was then dried to aconstant weight. A spectrophotometric correction forblood volume was included, since the volume of



ownloaded from


Laine and Allen Left Ventricular Edema, Fibrosis, and Function 1715

blood, and consequently water, found within thecoronary vasculature may vary significantly through-out the course of an experiment.7,23 Total tissuewater content is comprised of vascular water, inter-stitial water, and cellular water. Each of these watervolumes may be accurately determined. Total watercontent was obtained from the EVF of the tissuesample. Vascular water volume was obtained bothspectrophotometrically, using the technique ofPearce et al,23 and by chromium-51 labeling of redblood cells combined with hematocrit determina-tions.24 Interstitial fluid volume was determined us-ing technetium-99m-labeled diethylenetriaminepen-taacetic acid as an extracellular marker.24Subtraction produces cellular water volume. Four-micron sections were cut from samples of midmyo-cardium and stained (hematoxylin-eosin). Histologi-cal examination of these tissue samples was

performed to determine if myocytes were swollenfrom water accumulation, as seen in ischemic myocar-dium.25 A determination of hydroxyproline concentra-tion within the myocardium was carried out using themethod described by Weber et al12 and confirmed withan amino acid analyzer. Since collagen contains-13.4% hydroxyproline, the collagen content of thespecimen was calculated by multiplying the hydroxypro-line content by 7.46. The concentration of collagen wasexpressed as milligrams of collagen per 100 mg dryventricular weight. Protein concentration within thelymph draining the myocardial interstitium was deter-mined in each dog by use of a refractometer.15

Experimental ProtocolEach protocol was carried out for 3 hours before

determining the amount of myocardial edema or EVFpresent in the left ventricle. The various groups usedare summarized in Tables 1 and 2. Eight normotensivecontrols were used to determine the baseline EVF ofthe canine left ventricular myocardium.

Myocardial tissue fluid exits the heart primarily inthe form of lymph. Since lymph normally flows intothe superior vena caval system, elevation of SVCPreduces QLV.21 To determine the effect of reducingQLV alone for 3 hours, SVCP was elevated to 20mm Hg in three dogs (Fogarty balloon inflation). In12 dogs, myocardial microvascular pressure was ele-vated by inflation of the Swan-Ganz balloon catheterwithin the coronary sinus, which elevated CSP. Thiswas done since the rate at which fluid enters themyocardial interstitial space can be increased byelevation of pressure within the ventricular microvas-cular exchange vessels. In these dogs, cardiac func-tion curves were generated before and after 3 hoursof CSP elevation. In 13 dogs, QLV was reduced andCSP was elevated simultaneously. Cardiac functioncurves were also generated for these dogs before andafter the combined reduction of QLV and elevation ofCSP. Questions have been raised as to whetherincreases in CSP, such as those in our preparation,could produce a decrease in coronary artery flow thatwould result in compromised cardiac function. We

TABLE 1. Extravascular Fluid and Lymph Flow From the LeftVentricular Myocardium After Various Edemagenic Stresses

QLV EVFGroup n (ml/hr) at 3 hrs

Normotensivecontrols 8 3.1+2.1 2.90±0.20

Reduced QLV 3 >0.1 2.95±0.07Elevated CSP 12 29.5±7.3 3.45±0.16*tReduced QLV and

elevated CSP 13 >0.1 3.90±0.19*ttValues are mean±SD. n, Number of hearts; QLV, myocardial

lymph flow; EVF, extravascular fluid, expressed as the blood-free(wet weight-dry weight)/dry weight ratio; CSP, coronary sinuspressure.

*Cardiac function (from cardiac output vs. left atrial pressurecurves) significantly compromised vs. normotensive controls.

tSignificantly greater than normotensive controls.tSignificantly greater than CSP elevation alone.

have not been able to demonstrate a sustained de-crease in coronary artery flow after CSP elevation.This observation has recently been confirmed byWard et al.26 Baseline left ventricular EVF andcardiac function were determined in five dogs withchronic right ventricular hypertension. Control EVFand cardiac function data for dogs with chronicarterial hypertension were obtained in three dogs;four hypertensive dogs were exposed to elevated CSPfor 3 hours. End-diastolic left ventricular interstitialfluid pressure was determined in control and hyper-tensive dogs at baseline and after CSP elevation.Cardiac lymph samples were obtained from all dogsand underwent analysis for plasma protein concen-trations. At the conclusion of each 3-hour protocol,dogs were killed with an overdose of thiopentalsodium and 20 ml saturated potassium chloride solu-tion injected intravenously. The left ventricle (withoutthe septum) was removed in a consistent manner fromeach heart. Since the absolute value for hydroxy-proline concentration may vary, both as a function ofspecies and sampling site,27 full thickness samples ofmyocardium were removed from a consistent site inthe middle of the ventricle. The remaining ventricularmuscle was evaluated for the presence of edema bydetermining EVF as previously described.

Statistical AnalysisAll data are presented as mean±SD. Data analysis

was carried out on an IBM 4381 using the SASsoftware package.28 Groups of data were comparedby analysis of variance and F test. After analysis ofvariance, specific groups were tested using a t statisticwith a Bonferroni correction for multiple compari-sons.29 A value ofp<0.05 was considered significant.

ResultsData from a total of 48 dogs are presented. Table

1 summarizes the myocardial edema data from 36dogs. Myocardial EVF was not significantly elevatedafter 3 hours of reduced QLv alone. Elevation of CSPand the resultant increase in coronary microvascular



ownloaded from



TABLE 2. Extravascular Fluid and Interstitial Fluid Pressure ofthe Left Ventricular Myocardium After Chronic Right HeartPressure Elevation and Chronic Arterial Hypertension

End-diastolicleft ventricularinterstitial fluid

pressure EVFGroup n (mm Hg) at 3 hrs

Normotensive controls 8 14.9+3.1 2.90+0.20Chronic right heart

pressure elevation 5 ... 3.50+0.08*tChronic arterial

hypertension controls 3 24.8+3.7t 3.50+0.30* tChronic arterial

hypertension andacutely elevated CSP 4 47+7t- 4.20+--0.151ttValues are mean± SD. n, Number of hearts; EVF, extravascular

fluid, expressed as the blood-free (wet weight-dry weight)/dryweight ratio; CSP, coronary sinus pressure.

*Cardiac function (from cardiac output vs. left atrial pressurecurves) significantly compromised vs. normotensive controls.

tSignificantly greater than normotensive controls.*Significantly greater than hypertensive controls.

pressure caused significant edema formation over the3-hour protocol (EVF, 3.45+0.16). The combinedreduction of QLV and elevation of CSP also resultedin edema formation that was greater than the EVFfound after CSP elevation alone (EVF, 3.90±0.19).QLVunder control conditions and during each of theexperimental interventions is also shown on Table 1.

Table 2 summarizes the data from 12 additionaldogs and the eight normotensive controls. Baselineor control data are presented for three groups: 1)normotensive dogs, 2) dogs with chronic right heartpressure elevations, and 3) dogs with chronic arterialhypertension. Dogs with right-sided pressure eleva-tions had significantly more left ventricular edemathan did normotensive controls. The EVF of controldogs with chronic arterial hypertension was alsosignificantly greater than the EVF for normotensivecontrols. After elevation of CSP in dogs with arterialhypertension, EVF was significantly elevated abovethat in both the normotensive and hypertensive con-trols. Control left ventricular interstitial fluid pres-sure at end diastole was 14.9±3.1 mm Hg.15 Controlmyocardial interstitial fluid pressure increased to24.8±3.7 mm Hg when EVF rose to 3.5 ±0.30 inchronic arterial hypertension. Interstitial fluid pres-sure continued to increase to 47±7 mm Hg whenEVF rose to 4.20±0.15 in hypertensive dogs withCSP elevation. We found that hypertensive dogs withCSP elevation were prone to cardiac arrest whenstressed. Consequently, the volume infusions neces-sary to generate cardiac function curves were notadministered to this group.

Figure 1 demonstrates how SVCP, CSP, and QLVvaried throughout an experimental protocol. Addi-tional time was added to the beginning of thisprotocol so that variations in QLV could be seen afterchanges in CSP and SVCP. As CSP was elevated andtransmicrovascular fluid flux increased, QLV was seen

3TIME(hrs)

FIGURE 1. Time plot showing that after elevation of coro-nary sinus pressure (CSP), transmicrovascular fluid flux andleft ventricular lymph flow (QLV) increased. When superiorvena cava pressure (SVCP) is elevated, QLv can be totallyarrested.

4

to accelerate as well. When SVCP was elevated, thepressure that lymph must overcome in order to flowincreased, and QLV decreased to <1% of control.After elevation of CSP and SVCP, the fluid thatwould normally be leaving via the myocardial lym-phatics accumulates as myocardial edema.

Figure 2 presents the data for cardiac outputplotted as a function of left atrial pressure for a singledog. All control cardiac output curves were obtainedat the beginning of each protocol, and a seconddetermination was made after the accumulation ofmyocardial edema. In Figure 2, a myocardial EVF of3.9 can be seen to compromise cardiac function overa range of left atrial pressures. As seen in Table 1.cardiac function curves were significantly compro-mised after elevation of CSP alone for 3 hours. WhenQLV was reduced and CSP was elevated, cardiacfunction curves were further depressed. Cardiacfunction curves were also compromised in both thecontrol right heart pressure elevation group and thecontrol arterial hypertension group (Table 2).

Figure 3 represents the data from all dogs in whichcardiac output and EVF were determined. Cardiacoutput at a left atrial pressure of 15 mm Hg is plottedas a percent of control and as a function of EVF ormyocardial edema. This figure demonstrates the sig-nificant decrease in the heart's ability to maintaincardiac output at a constant left atrial pressure as

50-

40-

30-CSP(mmHg)20-

10-

0-



ownloaded from


Laine and Allen Left Ventricular Edema, Fibrosis, and Function

100

90-

.50h-

0Q

0CD

I

EEU,

11[L

0.

m

0

C)

0.

0

80

70

60

50

40

30-

20

10.

04 X3

5 10 15 20

LEFT ATRIAL PRESSURE (mm Hg)

FIGURE 2. Cardiac output plotted as a function of left atrialpressure. EVF, extravascular fluid expressed as the (wetweight-dry weight)Idry weight ratio. Edematous myocardiumwith an EVF of 3.9 exhibits a compromised cardiac functioncurve.

myocardial edema increases within the interstitium.Left ventricular end-diastolic pressure was signifi-cantly increased in chronic dogs with elevated EVFs(from 4.5+ 1.0 to 15.2±2.5 mm Hg). Heart rate wassignificantly elevated in these halothane-anesthetizeddogs (147±12 beats/min) and could not be shown tochange after elevation of left atrial pressure orinterstitial fluid accumulation.

Table 3 provides a comparison between variousleft ventricular parameters in control dogs and inthose with chronic right heart pressure elevations.Left ventricles (without septum) were weighed afterremoval of a consistent full-thickness slice of myocar-dium for histological examination. No statistical in-crease could be demonstrated in the wet mass be-tween the two groups. Left ventricular dry weightsfrom the two groups, 12.45+0.85 g (normotensivecontrols) and 12.57±0.90 g (dogs with chronic rightheart pressure elevation), also showed no difference.Dogs with chronic arterial hypertension had a leftventricular dry weight of 17.23±1.10 g, which is 28%greater than that for the control and right heartfailure groups. We believe this statistically significantincrease in left ventricular mass results from leftventricular hypertrophy. The percent blood content

EXTRAVASCULAR FLUID (WD)

FIGURE 3. Cardiac output at a left atrial pressure (LAP) of15 mm Hg, expressed as percent of control, plotted as a

function of extravascular fluid. W, wet weight; D, dry weight.As myocardial edema accumulates, the heart's ability tomaintain cardiac output at LAP of 15 mm Hg diminishes.

of ventricles from each group was not statisticallydifferent. The percentage of the left ventricle that iswater and the percent of the left ventricle that isinterstitium increased significantly in both the right-sided pressure elevation and chronic hypertensionmodels. The percent of the left ventricle that isinterstitium was determined using the extracellularmarker technetium-99m-labeled diethylenetriamine-pentaacetic acid. Knowing the total water volume inthe left ventricular myocardium and subtractingblood volume and interstitial volume, we were unableto demonstrate any change in myocyte volume in thechronic right heart pressure elevation model. Thiswas consistent with our histological findings. Thequantity of collagen within the left ventricle in-creased significantly both in the right heart failuremodel (5.80+0.28) and in the chronic arterial hyper-tension model (5.1+0.6) when compared with con-trols (3.90±0.64). This is consistent with the findingsof others.30 The lymph/plasma protein concentrationratio was significantly decreased in the chronic rightheart pressure elevation model (from 0.83+± 0.05 to0.70+0.03) and increased to 0.88+0.05 in the arterialhypertension model. These changes are consistentwith a decrease in microvascular permeability in the

12.0

11.0*

10.0'

9.0'

8.0

7.0

6.0

5.0

4.0

c

CL

00

0

./ %

//

//

//

//

A(3.0-

2.0-

1.01

1'

.. * CONTROLA EVF = 3.9

n . ,

1717

,^",



ownloaded from



TABLE 3. Comparison of Left Ventricular Parameters for Control Dogs and Those With Chronic Right HeartPressure Elevation

LV LVLV mass blood LV H20 interstitium LV collagen LV lymph

Group n (g) (%) (blood-free) (%) (%) (g%) (CL/CP)Normotensive controls 7 57.1+15.1 12.1+6.1 75.2+0.5 17+3 3.90+0.64 0.83+0.05Chronic right heart pressure

elevation 5 67.9+11.6 13.5+4.7 78.6+0.7* 26±4* 5.80±0.28* 0.70+0.03*

Values are mean--+SD. n. number of dogs. LV, leftconcentration ratio.

*Significantlv greater than normotensive controls.

right heart failure (low-protein) model and an in-crease in microvascular permeability in the chronicarterial hypertension (high-protein) model.

DiscussionWe studied the relation between left ventricular

edema, myocardial interstitial fibrosis, and cardiacfunction. The sequential series of events depicted inFigure 4 provides an orderly framework around whicha discussion of these relations may be structured.

Myocardial Transmicrovascular Fluid FluxMyocardial edema originates as fluid that crosses

the microvascular exchange vessels and enters theinterstitial space. The factors that govern fluid move-ment into the myocardial interstitium and thus con-trol myocardial tissue fluid volume may be expressedin the form of the Starling equation31:

Jv=Kf[(Pc-Pint) -Cr(7rC- Trint)I (1)

where Jv is the total volume of fluid crossing themicrovascular exchange barrier and entering themyocardial tissue spaces,20 Kf is the filtration coeffi-cient; PC and Pm, are hydrostatic pressures within themicrovascular exchange vessels and within the inter-stitial space, respectively; o- is the reflection coeffi-cient; and ir. and Tin,t are colloid osmotic pressureswithin the microvessels and within the interstitium,respectively. Normally, a volume of fluid equal to Jvis removed from the heart via the myocardial lym-

ventricular myocardium; CL/CP, lymph/plasma protein

phatics (QLv) and via transudation across the surfaceof the heart into the pericardial sac; thus, the heartdoes not become edematous. A majority of myocar-dial tissue fluid exits the heart via the lymphaticsystem. As such, any reduction in LV should poten-tiate edema formation. Since the lymphatics draininto the central venous circulation, any elevation invenous pressure tends to decrease QLV, as seen inFigure 1. In the present study, elevation of SVCP anddecreased LV alone did not produce significant leftventricular edema formation over the 3-hour timecourse of the experimental protocol. This is consis-tent with findings in other organs that longer timeperiods may be necessary to generate sufficient vol-umes of edema for quantitation and that otherfactors within the Starling equation referred to as"defense mechanisms against edema formation"tend to protect the heart against fluid accumulationduring acute insults.32 Kf is the filtration coefficientand is a function of the hydraulic conductivity of themicrovascular exchange barrier as well as the surfacearea over which water exchange may take place. P, isthe hydrostatic pressure within the microvascularexchange vessels and acts to force fluid into themyocardial interstitium. When coronary venous pres-sures (coronary sinus and thebesian veins) wereacutely elevated, P, was elevated, and we foundincreased QLV (Figure 1) and significant left ventric-ular myocardial edema formation. This edema for-mation was significantly exacerbated by simultaneous

ELEVATED

CENTRAL DECREASEDVENOUS MYOCARDIAL

PRESSURE LYMPH FLOW

~~+

RIGHT ELEVATEDHEART CORONARY

FAILURE VENOUS

PRESSURES

CHRONIC INCREASEDARTERIAL _ MICROVASCULAR

HYPERTENSION PERMEABILITY

MYOCARDIALEDEMA

OR

INCREASEDINTERSTITIALFLUID VOLUME

ALTERED COPMIEMYOCARDIAL COMPROMISEDINTERSTITIAL CARDTIAC

MATRIX ~ FNTO

(e g COLLAGEN.COMPLIANCE 1

FIGURE 4. Flow chart depicting several mechanisms that potentiate myocardial edema formation. Myocardial edema maycompromise cardiac function directly or may alter the interstitial matrix, thus indirectly compromising function.



ownloaded from



reduction of QLV. The synergistic relation betweenelevated coronary venous pressures and decreasedQLV potentiates left ventricular edema formation andis represented by the plus sign in Figure 4. It shouldbe noted that our chronic right heart failure modelincreases the pressure experienced by the left ven-tricular coronary microvasculature both by elevatingpressure to the thebesian veins draining into the rightventricle and, after fluid retention and redistribution,by elevating CSP. The increase in circulating fluidvolume during right heart failure elevates centralvenous pressure and potentiates left ventricularedema formation through the "decreased myocardiallymph flow" mechanism as well (Figure 4).

Pin, is the hydrostatic pressure within the intersti-tial space and, as it increases, functions to opposefluid movement into the interstitium. This increase ininterstitial pressure is one of the factors that preventssignificant edema formation after reduction of QLValone over the short time course of the acute studies.17c and Trjint are the colloid osmotic pressures withinthe microvessels and interstitium, respectively. Thereflection coefficient (o-) is an index of microvascularpermeability. Increases in microvascular permeability(or decreases in cr) allow additional plasma filtrate tocross the microvascular exchange barrier at normalpressures and to form high-protein edema. De-creases in microvascular permeability restrict plasmafiltration and produce low-protein edema.

In the present study, we found that the increasedpermeability in chronic hypertension produced an ele-vated baseline level of left ventricular myocardial inter-stitial fluid, which was significantly exacerbated afterelevation of CSP. It is important to note that the leftventricular edema levels found in both chronic arterialhypertension and chronic right heart pressure elevation(EVF, -3.5) are not apparent without careful quanti-tation and may be easily overlooked. As seen in Figure5, a change in myocardial edema from control to 3.5represents only a 3% change in myocardial watercontent and -12% change in heart weight.

Myocardial Edema and Interstitial FibrosisThe fact that myocardial interstitial fibrosis can

compromise cardiac function has been well docu-mented and is portrayed schematically in Figure 4.12,33Our primary concern was with the deposition ofmyocardial interstitial collagen stimulated by the pres-ence of myocardial edema. Since interstitial fibrosis isa common sequela in response to the presence ofinterstitial edema,34 we were not surprised to find thatleft ventricular myocardial collagen concentrationrose in response to the chronic edema associated witharterial hypertension and right heart pressure eleva-tions (Table 3). We had speculated that the fibroticchanges seen in chronic arterial hypertension resultednot only from myocardial remodeling during hypertro-phy but also from the presence of myocardial edema.Although we could detect significant volumes ofedema in our hypertensive preparations, we wereunable to differentiate fibrosis secondary to hypertro-

100-

90-

80

C

0

0 ax

_ >

O'

I-zW

C.)

70-

60-

50-

40-

30-

20-

10-

0!9.

/

I//

/IIII/I/II

/,

23 4 5 6

100

90

80I

O3:X_i:70

60

>-_O a,

50

40 Z-IQ

30 Z0trW

20 X

10

0

WET Wt - DRY Wt EVFDRY Wt

FIGURE 5. Graphic representation of how the percent waterwithin the myocardium and the percent change in totalmyocardial weight vary at different levels ofextravascularfluid(EVF).

phy from that resulting from the presence of edema.Therefore, we began using a model of chronic myo-cardial edema in which no left ventricular myocardialhypertrophy could be demonstrated (Table 3). Thismodel of pure myocardial edema was produced bychronically banding the pulmonary artery. In thispreparation, central venous pressure is elevated, thusreducing lymph flow from the myocardial lymphaticsand potentiating edema formation. Elevation of pres-sure within the coronary sinus and thebesian veins alsocaused an acceleration of fluid movement into themyocardial interstitium, thus further potentiating leftventricular interstitial fluid accumulation. Since theleft ventricle in this preparation does not pumpagainst an elevated pressure, no left ventricular hyper-trophy takes place. We believe that a majority of thefibrosis seen in this preparation is reactive and not areparative response to tissue damage.35 In this modelof myocardial edema, significant collagen depositionwas found at edema volumes similar to those found inthe chronic arterial hypertension preparation.We may speculate as to what signal is transduced

by the fibroblasts to potentiate collagen gene activityin the presence of interstitial edema.36 We shouldnote that not only the type and quantity of collagenbut the manner in which it is deposited or organizedcan vary significantly.37'38 Many investigators believe



ownloaded from



that the presence of high-protein concentration withinedema stimulates the deposition of fibrotic material.34We believe that the protein concentration of the edemamay influence the type of collagen deposited but notwhether matrix will be formed, since we have seencollagen deposition in both a high-protein edemamodel (chronic arterial hypertension) and a low-pro-tein edema model (right heart failure).

It has been demonstrated that collagen depositionmay be stimulated in fibroblasts that are exposed toelevated pressure.39 Since the accumulation of myo-cardial interstitial edema increases interstitial fluidpressure (Table 2), fibroblasts may respond to thiselevated pressure by increasing the strength of thecollagen superstructure of the heart via fibrous dep-ositions. It has also been speculated that compro-mised oxygen delivery, from reduced blood flow, andrelative ischemia within the myocardium may lead tointerstitial fibrosis.40 Ischemia may also result fromincreased interstitial diffusion distances between thecapillaries and myocytes resulting from accumulationof edema (Table 3).41,42 In either case, the coronaryangiogenesis associated with ischemia or increasedwork load results in the deposition of new tissuecomponents.43 The evaluation of which factors asso-ciated with edema formation stimulate collagen dep-osition remains an active area of investigation. Weshould point out that, when chronic arterial hyper-tension is brought under control or systemic venouspressure is lowered in our chronic right heart failuremodel, edema regresses very slowly, and the elevatedcollagen concentrations and compromised cardiacfunction curves (relative to control) remain for ex-tended periods of time.44

Myocardial Edema and Cardiac FunctionAs expressed schematically in Figure 4 and graph-

ically in Figures 2 and 3, both acute and chronicmyocardial edema compromise cardiac function. Aspointed out in the previous section, cardiac functionis also compromised by the deposition of interstitialfibrosis associated with chronic edema. An increasein myocardial interstitial fluid from a control EVFvalue of 2.9 to 3.5, as seen during chronic arterialhypertension and chronic right heart pressure eleva-tions, represents a change from a control myocardialwater content of 75% to 78%. This small change inthe percentage of the myocardium that is water signif-icantly compromises function. This same change ac-counts for a 12% increase in the total weight of theheart (Figure 5). Such changes in myocardial watervolume account for a decrease of -30% in the heart'sability to maintain cardiac output at a left atrial pres-sure of 15 mm Hg (Figure 3). As with other factors thatcan compromise cardiac function, compromised func-tion due to myocardial edema formation manifestsitself as a decrease in cardiac reserve and an inability ofthe heart to perform when stressed. When the EVFvalue increased to -4, the amount of water within themyocardial tissue only increased by 5%, and total heartweight increased by 25%. As seen in Figure 3, this

increase in interstitial fluid volume resulted in morethan a 50% decrease in the heart's ability to maintainthe elevated cardiac output that the controls coulddevelop.As we have demonstrated in normal hearts, eleva-

tion of CSP alone requires several hours to producesufficient edema to quantitate and to compromisecardiac function.7 Acute elevations in healthy sub-jects, for a period of minutes, will not producesignificant myocardial edema or compromise cardiacfunction.26 In contrast, elevation of CSP in subjectswith chronic disease, such as elevated systemic ve-nous pressure, produces rapid accumulation ofedema and compromised cardiac function.74>

Several mechanisms may account for the compro-mised cardiac function seen in the presence of myo-cardial edema. We believe that all of the followingmechanisms may in fact act in concert to compromisefunction. As myocardial edema accumulates withinthe interstitial spaces, interstitial pressure rises, thusincreasing the stiffness of the myocardium. Thisstiffness combined with the viscous effects of movingexcess interstitial water on a beat-to-beat basis cancompromise the heart's ability to contract efficient-ly.46 Although interstitial collagen has great tensilestrength, increases in interstitial volume and pressuremay displace the collagen fibers and potentially un-couple or break collagen struts loose from theiranchoring points on fibroblasts.33,47 Because of theheart's reliance on a well-organized interstitial matrixaround which to contract, a disruption in the collagenstructure could also compromise function. As myo-cardial edema accumulates within the interstitium,the diffusion distance for oxygen to the myocytesincreases. This is of particular importance in anorgan such as the heart, which operates near maxi-mum oxygen extraction capacity at all times. Recentreports document that myocardial infarctions growmore rapidly in the hearts of subjects with chronicarterial hypertension.42 These reports speculate thatincreased diffusion distances for oxygen and relativeischemia,41 due to the presence of myocardial edema,may exacerbate the rapid infarction growth.We conclude that acute myocardial edema com-

promises cardiac function and that chronic rightheart failure and chronic arterial hypertension alsoproduce left ventricular myocardial edema, whichcompromises function in these common models ofinterstitial heart disease. The presence of myocardialedema in these chronic models also potentiates in-terstitial fibrosis, leading to a further decrease in theheart's ability to function normally.

AcknowledgmentsWe wish to thank Sally Eustice and Marie Steen-

berg for their excellent technical assistance and Car-oline Buggs-Warner for the typing of this manuscript.

References1. Cross C, Rieben P, Salisbury P: Influence of coronary perfu-

sion and myocardial edema on pressure-volume diagram ofleft ventricle. Am J Physiol 1961;201:102-108



ownloaded from



2. Pogatsa G, Dubecy E, Gabor G: The role of myocardial edemain the left ventricular diastolic stiffness. Basic Res Cardiol1976;71:263-269

3. Amano J, Sunamori M, Kameda T, Okamura T, Suzuki A:Correlation among water content, left ventricular function,coronary blood flow, and myocardial metabolism after hypo-thermic ischemic cardiac arrest. Adv Myocardiol 1983;4:465-571

4. Sasaguri S, Sunamori M, Saito K, Suzuki A: Early change ofmyocardial water during acute cardiac allograft rejection. JpnCirc J 1986;50:1113-1119

5. Laks H, Standeven J, Blair 0, Hahn J, Jellinek M, Willman V:The effects of cardiopulmonary bypass with crystalloid andcolloid hemodilution on myocardial extravascular water. JThorac Cardiovasc Surg 1977;73:129-138

6. Laine G: Microvascular changes in the heart during chronicarterial hypertension. Circ Res 1988;62:953-960

7. Laine G: Change in (dP/dt)m,,. as an index of myocardialmicrovascular permeability. Circ Res 1987;61:203-208

8. Miller A: Pathology of interference with cardiac lymph flow, inLymphatics of the Heart. New York, Raven Press, Publishers,1982, pp 262-334

9. Symbas P, Cooper T, Gantuer G, Willman V: Lymphaticdrainage of the heart: Effects of experimental interruption oflymphatics. Surg Forum 1963;14:254-256

10. Kluge T, Ullal S: Pathology of the heart following chroniccardiac lymphatic obstruction. Acta Pathol Microbiol Scand [A]1972;80:150-158

11. McKinney B: Endocardial changes produced in Patus mon-

keys by the ablation of cardiac lymphatics and the administra-tion of a plantain diet. Am Heart J 1976;91:484-491

12. Weber K, Janicki J, Shroff S, Pick R, Chen R, Bashey R:Collagen remodeling of the pressure-overloaded, hypertro-phied nonhuman primate myocardium. Circ Res 1988;62:757-765

13. Weber K, Janicki J, Pick R, Abrahams C, Shroff S, Baskey R,Chen R: Collagen in the hypertrophied, pressure-overloadedmyocardium. Circulation 1987;75(suppl I):I-40-I-47

14. Michael LH, Lewis RM, Brandon TA, Entman ML: Cardiaclymph flow in conscious dogs.Am JPhysiol 1979;237(Heart CircPhysiol 6):H311-H317

15. Laine G, Granger H: Microvascular, interstitial, and lymphaticinteractions in normal heart. Am J Physiol 1985;249:H834-H842

16. Davis JO, Hyatt RE, Howell DS: Right-sided congestive heartfailure in dogs produced by controlled progressive constrictionof the pulmonary artery. Circ Res 1955;3:252-258

17. Yankopoulos NA, Davis JO, Cotlove E, Trapasso M: Mecha-nism of myocardial edema in dogs with chronic congestiveheart failure. Am J Physiol 1960;199:603 -608

18. Olivetti G, Ricci R, Lagrasta C, Maniga E, Sonnenblick EH,Anversa P: Cellular basis of wall remodeling in long-termpressure overload-induced right ventricular hypertrophy inrats. Circ Res 1988;63:648-657

19. Laine G, Granger H: Permeability of intestinal microvessels inchronic arterial hypertension. Hypertension 1983;5:722-727

20. Curry FE: Mechanics and thermodynamics of transcapillaryexchange, in Benkin EM, Michel CC (eds): Handbook ofPhysiology, Section 2: The Cardiovascular System, Volume IVWashington, DC, American Physiological Society, 1984, pp309-374

21. Laine G, Drake R, Allen S, Gabel J: Effect of systemic venouspressure elevation on lymph flow and lung edema formation. JAppl Physiol 1986;61:1634-1638

22. Laine G, Drake R, Zavisca F, Gabel J: Effect of lymphaticcannula outflow height on lung microvascular permeabilityestimation. JAppl Physiol 1984;57:1412-1416

23. Pearce M, Yamashita J, Beazell J: Measurement of pulmonaryedema. Circ Res 1962;16:482-488

24. Parker JC, Falgout HJ, Grimbert FA, Taylor AE: The effect ofincreased vascular pressure on albumin-excluded volume andlymph flow in the dog lung. Circ Res 1980;47:866-875

25. Dauber IM, VanBenthuysen KM, McMurtry IF, Wheeler GS,Lesnefsky EJ, Horwitz LD, Weil JV: Functional coronarymicrovascular injury evident as increased permeability due tobrief ischemia and reperfusion. Circ Res 1990;66:986-998

26. Ward K, Fisher D, Michael L: Elevated coronary sinus pres-sure does not alter myocardial blood flow or left ventricularcontractile function in mature sheep. J Thorac Cardiovasc Surg1988;95:511-515

27. Hance AJ, Crystal RG: The connective tissue of lung. Am RevRespir Dis 1975;112:657-711

28. Barr A, Goodnight J, Sall J, Helwiz J: A User's Guide toSAS-76. Raleigh, NC, SAS Institute Inc, 1976, pp 151-256

29. Neter J, Wasserman W: Applied Linear Statistical Models.Homewood, Ill, Irwin, 1974, pp 160-165

30. Williams JF Jr: Evolving concepts in congestive heart failure.Mod Conc Cardiovasc Dis 1990;59(pt 1):43-46

31. Starling EH: On the absorption of fluids from the connectivetissue spaces. J Physiol (Lond) 1896;19:312-326

32. Guyton AC, Barber BJ, Moffatt DS: Theory of interstitialpressures, in Hargens AR (ed): Tissue Fluid Pressure andComposition. Baltimore, Williams & Wilkins Co, 1981, pp11-19

33. Weber K, Pick R, Moe G, Zucker I, Armstrong P: Collagenand the failing canine ventricle (abstract). Circulation 1989;80(suppl II):II-504

34. Witte CL, Witte MH, Dumont AE: Pathophysiology of chronicedema, lymphedema, and fibrosis, in Staub NC, Taylor AE(eds): Edema. New York, Raven Press, Publishers, 1984, pp521-542

35. Jalil J, Janicki J, Pick R, Abrahams C, Weber K: Fibrosis-induced reduction of endomyocardium in the rat after isopro-terenol treatment. Circ Res 1989;65:258-264

36. Chapman D, Weber KT, Eghbali M: Regulation of fibrillarcollagen types I and III and basement membrane type IVcollagen gene expression in pressure overloaded rat myocar-dium. Circ Res 1990;67:787-794

37. Borg TK, Caulfield JB: The collagen matrix of the heart. FedProc 1981;40:2037-2041

38. Robinson TF, Factor SM, Capasso JM, Wittenberg BA,Blumenfeld 00, Seifter S: Morphology, composition, andfunction of struts between cardiac myocytes of rats andhamster. Cell Tissue Res 1987;249:247-255

39. Riley D, Gullo J: Pressure applied to cultured pulmonaryartery endothelial cells causes release of a fibroblast mitogenand induces a proto-oncogene (abstract). FASEB J 1988;2:A300

40. Hittinger L, Shannon R, Bishop S, Gelpi R, Vatner S:Subendomyocardial exhaustion of blood flow reserve andincreased fibrosis in conscious dogs with heart failure. Circ Res1989;65:971-980

41. Ziegler W, Goresky C: Transcapillary exchange in the workingleft ventricle of the dog. Circ Res 1971;29:181-207

42. Dellsperger K, Clothier J, Hartnett J, Haun L, Marcus M:Acceleration of the wavefront of myocardial necrosis bychronic hypertension and left ventricular hypertrophy in dogs.Circ Res 1988;63:87-96

43. Tomanek R, Schalk K, Marcus M, Harrison D: Coronaryangiogenesis during long-term hypertension and left ventricu-lar hypertrophy in dogs. Circ Res 1989;65:352-359

44. Stacy D, Prewitt R: Effects of chronic hypertension and itsreversal on arteries and arterioles. Circ Res 1989;65:869-879

45. Ilbawi M, Idriss F, Muster A, DeLeon S, Berry T, Duffy CPaul M: Effects of elevated coronary sinus pressure on leftventricular function after the Fontan operation. J ThoracCardiovasc Surg 1986;92:231-237

46. Weng Z, Schierman S, Goldstein A, Nicolosi A, Spotnitz H:Effects of perfusate tonicity on myocardial edema and com-pliance (abstract). Circulation 1988;78(suppI II):II-70

47. Capasso JM, Robinson TF, Anversa P: Alterations in collagencross-linking impair myocardial contractility in the mouseheart. Circ Res 1989;65:1657-1664

KEY WORDS * interstitial heart disease * interstitial matrixinterstitial fluid pressure * cardiac lymph collagen depositiona hypertension * right heart failure



ownloaded from


G A Laine and S J AllenLeft ventricular myocardial edema. Lymph flow, interstitial fibrosis, and cardiac function.

Print ISSN: 0009-7330. Online ISSN: 1524-4571 Copyright © 1991 American Heart Association, Inc. All rights reserved.is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Circulation Research

doi: 10.1161/01.RES.68.6.17131991;68:1713-1721Circ Res.

http://circres.ahajournals.org/content/68/6/1713World Wide Web at:

The online version of this article, along with updated information and services, is located on the

http://circres.ahajournals.org//subscriptions/

is online at: Circulation Research Information about subscribing to Subscriptions:

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

document. Permissions and Rights Question and Answer about this process is available in the

located, click Request Permissions in the middle column of the Web page under Services. Further informationEditorial Office. Once the online version of the published article for which permission is being requested is

can be obtained via RightsLink, a service of the Copyright Clearance Center, not theCirculation Researchin Requests for permissions to reproduce figures, tables, or portions of articles originally publishedPermissions:



ownloaded from

http://circres.ahajournals.org/content/68/6/1713

http://www.ahajournals.org/site/rights/

http://www.lww.com/reprints

http://circres.ahajournals.org//subscriptions/


left ventricular myocardial edema - circulation...

Documents