A Comparison of the Effects of Vasodilator Stimuli

on Peripheral Resistance Vessels in Normal Subjects

and in Patients with Congestive Heart Failure

ROBERTZEuIs, DEANT. MASON,and EUGENEBRAUNWALDwith the technical assistance of MIMI WINTHALTR2 and CLARAKING

Cardiology Branch, National Heart Institute, National Institutes of Health,Bethesda, Maryland 20014

A B ST R A C T The objective of this investigationwas to characterize the mechanism of peripheralvasoconstriction observed in heart failure and todetermine whether it can be attributed to the aug-mented sympathetic nervous activity, characteristicof this state. The response of the resistance bed inthe forearm after release of inflow occlusion(reactive hyperemia), to hand exercise, and tolocal heating and the response of the calf resistancevessels to arterial occlusion and intrararterial so-dium nitrite and phentolamine were studied in 23patients with congestive heart failure and 21 nor-mal subjects. In the normal subjects, reactivehyperemia blood flow after varying periods ofarterial occlusion greatly exceeded the values ob-served in patients with heart failure. Local anes-thetic blockade and intra-arterial phentolamine didnot significantly alter the reactive hyperemia re-sponse in heart failure patients, militating againstthe possibility that increased sympathetic vaso-constrictor activity is responsible for the reductionof this response. Following compensation, thereactive hyperemia response returned toward nor-mal. The striking elevations of the forearm bloodflow observed after hand exercise and heating ofthe forearm in normal subjects were also markedlyattenuated in patients with heart failure. Followingintra-arterial phentolamine and/or sodium nitrite,peak calf blood flow was still significantly reducedin heart failure.

Received for publication 5 September 1967 and in re-vised form 23 October 1967.

These observations indicate that (1) heart fail-ure is characterized by a striking reduction in theresponse to a variety of endogenous and exogenousvasodilator stimuli; (2) circulating catecholaminesand sympathetic vasoconstrictor activity are notsolely responsible for the elevation of peripheralvascular resistance and the reduced response tovasodilator stimuli in heart failure; and (3) heartfailure may increase systemic vascular resistancedirectly by altering the mechanical properties andreducing the dilating ability of the resistancevessels.

INTRODUCTION

It is generally acknowledged that abnormalities ofthe peripheral circulation occur commonly in pa-tients with congestive heart failure; blood flow tomany organs has been shown to be reduced, andsince arterial pressure is generally normal, theresistance to blood flow through these vascularbeds is elevated. It has been thought that arteriolarconstriction, mediated by the sympathetic nervoussystem, is responsible for these elevations of vascu-lar resistance and that this peripheral vasoconstric-tion is important in maintaining the perfusionpressure to vital organs, such as the heart andbrain, in the face of a low cardiac output. How-ever, it is clear that vascular resistance is deter-mined not only by nervous influences but is af-fected profoundly by local metabolic stimuli as wellas by the physical characteristics of the blood

960 The Journal of Clinical Investigation Volume 47 1968

vessels themselves. It is not known whether thealterations in the peripheral vasculature in con-gestive heart failure are due entirely to alteredautonomic influences or whether abnormalities inthe response of the vessels to metabolic vasodilatorstimuli also play a role. Accordingly, the objectiveof the present investigation was to compare theeffects of a variety of endogenous and exogenousvasodilator stimuli on normal subjects and pa-tients with congestive heart failure.

METHODSStudies were carried out in a total of 21 normal subjectsbetween 19 and 50 yr of age (average, 27.3 + 2.3 yr/SEM)and 23 patients with congestive heart failure between 23and 56 yr of age (average, 43.8 + 1.7 yr). All of the pa-tients in the latter group were receiving digitalis glyco-sides. 21 patients had rheumatic heart disease, one hadcor pulmonale, and one had idiopathic hypertrophic sub-aortic stenosis. All of these patients had hepatomegaly,venous pressures exceeding 15 cm H20, and were infunctional Classes III or IV, according to the New YorkHeart Association. Peripheral edema was present in allpatients in whom forearm blood flow was measured, butwas absent in all patients in whom calf blood flow wasmeasured. Mean arterial pressure averaged 82.5 ± 1.9mmHg in the normal subjects and 87.9 + 2.5 mmHg inthe patients with congestive heart failure. Forearm andcalf blood flow were measured with the limbs at heartlevel and the subject recumbent. The venous occlusiontechnique was employed (1, 2) utilizing a single strandmercury-in-rubber strain gauge plethysmograph as previ-ously described by Holling, Boland, and Russ who vali-dated the use of this method (3). Circulation to the handor foot was arrested by inflating a cuff around the wristor ankle for at least 1 min before each determination ofblood flow (4). All measurements were performed withthe subjects in the basal, postabsorptive state, with atleast a 15 min period of equilibration being allowed toelapse after placement of all the instruments before bloodflow recordings. Arterial pressure was measured in theopposite brachial artery through an indwelling needle orby the auscultatory method. Regional vascular resistancewas calculated by dividing the difference between themean arterial pressure and venous pressure by the bloodflow. Minimal vascular resistance was estimated usingthe peak blood flow after release of 10 min of arterialocclusion.

I. Endogenous Vasodilator Stimuli

A. REACrIVE HYPEREmiAControl blood flow was taken as the average of 6-10

flow measurements made at 15-sec intervals before eachintervention. Ischemia of the forearm or calf was pro-duced by inflating the upper cuff to suprasystolic pres-sure for 1, 3, 5, and 10 min. After release of arterial

occlusion, blood flow was measured initially at 5 and 15sec after release and every 15 sec thereafter until it re-turned to control values. Reactive hyperemia blood flow(RHBF), Fig. 1A, was defined as the total excess bloodflow in ml/100 g above the control resting flow.

Three interventions were utilized to attempt to alterthe reactive hyperemia response by modifying peripheralsympathetic tone. In the first, reactive hyperemia after5 min of ischemia was studied before and after 1.0 mg ofphentolamine was injected into the femoral artery. Ar-terial occlusion was applied 3 min after injection of thedrug at a time when the blood flow had stabilized at itsnew resting level. Alpha adrenergic blockade with thisdose of phentolamine was insured by observing markedreduction of the constrictor response to intra-arterialnorepinephrine and to immersion of the hand in ice water.In the patients with heart failure, before the administra-tion of phentolamine, the injection of 1 ,ug of norepi-nephrine and application of the cold pressor test reducedcalf blood flow by averages of 44 and 26%o, respectively;after phentolamine these responses were significantlyless, averaging 16 and 5%o, respectively. The second in-tervention consisted of observing the reactive hyperemiaresponse of the forearm to 1, 3, and 5 min of arterialocclusion before and after blocking the median, ulnar,and radial nerves at the elbow with xylocaine. The pres-ence of sympathetic blockade was confirmed by observingthat resting blood flow increased and that the vasocon-strictor response to cold stimulation was abolished. Thiswas documented by the finding in the patients with heartfailure that before nerve blockade, forearm blood flowfell by an average of 35% during the cold pressor test.Nerve blockade resulted in a tripling of resting fore-arm blood flow and an actual reversal in the blood flowresponse to the cold pressor test; a 28%o increase wasobserved, presumably as a consequence of reflex elevationof systemic arterial pressure. Third, norepinephrine wasinfused i.v. at a rate of 16 ,ug/min to elevate mean ar-terial pressure by 20-30 mmHg. Circulating catechola-mine concentrations were determined by the trihydroxy-indole method (5). The reactive hyperemia response wasstudied before and during the norepinephrine infusion.

Venous congestion was produced by the method de-scribed by Wood, Little, and Wilkins (6), in which thearm was packed with blood by inflating the upper armcuff to 70-95 mmHg for 5 min before 10 min of arterialocclusion. The reactive hyperemia response was thenmeasured.

The reactive hyperemia response was studied in fivepatients with heart failure shortly after admission andafter compensation.

B. AcrIVE HYPEREMIA

The forearm was exercised by squeezing a hand dy-namometer to 50 lb. pressure, once every second for 30sec. The flow was recorded at 5 and 15 sec after the finalcontraction and then every 15 sec until it returned tocontrol levels. The postexercise hyperemia in ml/100 gwas calculated as the total blood flow in excess of restingblood flow (Fig. 1C).

Vasodilator Stimuli in Heart Failure 961

C. DIRECT HEATING

The upper forearm was heated by means of two 150 wtungsten lamps with internal reflectors, placed 30 cmfrom the arm and 1800 from one another. The straingauge was protected by an opaque shield which did nottouch the arm. Comparisons were made on the basis ofthe total excess blood flow above control flow, observedduring 5 min of heating and during the following 5 minof recovery (Fig. 1E).

II. Exogenous Vasodilator Stimuli

A. VASODILATORDRUGS

Sodium nitrite (NaNO2), 10, 30, and 50 mg, was in-jected through an indwelling Cournand needle in the. fe-moral artery, and calf blood flow recorded every 15 secuntil the peak effect had subsided. 15 min was allowedto elapse between injections. After the effects of theNaNOa had been dissipated, phentolamine (1 mg) wasinjected intra-arterially and calf blood flow was recordedevery 15 sec. The peak response was noted, as well asthe new steady level of resting blood flow, which wasreached 3 min after the injection. This level was com-pared to the resting blood flow noted before adrenergicblockade. 5 min after the intra-arterial administration ofphentolamine, 30 mg of NaNO2was again injected intra-arterially and the calf blood flow response compared withthat produced by the same dose of sodium nitrite givenbefore phentolamine. Again the presence of sympathetic

NORMAL

blockade was confirmed as indicated above. Methemoglo-bin levels determined by the cyanide method (7) werefound not to exceed 5%o with these doses of NaNO2.

RESULTS

I. Basal vascular dynamics. The basal forearmblood flow determined in 14 normal subjects aver-aged 4.12 + 0.53 (SEM) ml/min per 100 g andwas significantly greater than that observed in 17patients with congestive heart failure, in whom itaveraged 2.09 + 0.19 ml/min per 100 g (P <0.01). Forearm vasuclar resistance averaged 24.0+ 1.5 mmHg per ml/min per 100 g in the normalsubjects and was significantly higher in the heartfailure patients, 47.5 + 5.1 mmHg per ml/minper 100 g (P < 0.01), Fig. 3, left. Similarly, thecalf blood flow of seven normal subjects was sig-nificantly higher (4.71 +- 0.35 ml/min per 100 g)than in six patients with congestive heart failure(2.93 + 0.56 ml/min per 100 g), P < 0.01,whereas vascular resistance in the calf averaged20.7 + 1.3 mmHg per ml/min per 100 g in nor-mal subjects and 33.6 ± 3.6 mmHg per ml/minper 100 g in patients with heart failure (P < 0.01).

II. Reactive hyperemia. The ability of the fore-

HEART FAILURED. Y 06-80-9/ REACTIVE HYPEREMIA FW 06-89-99

A B40 PEAK RHBF

45.7 ml / min / 100 g

30

ISCHEMIA (5min)20

20E,-PEAK RHBF9 ~~~~~~~~~~~20.0ml / min / 100 g

10 - TOTAL RHBF6.9m100 9g TOTAL RHBF

f ISCHEMIA (5rmin)8,

30ACTIVE HYPEREMIA

c~~~~~~~~20 -FOREARMEXERCISE

f FOREARMEXERCISE'OF E DIEC HEATING F.

HEAT ON OFF

HEAT ON OOFF0

0 2 4 6 8 10 0 2 4 6

MINUTES8 10

FIGURE 1 Serial forearm blood flowmeasurements in normal subject (left)and a patient with congestive heartfailure (right). A and B, blood flowrecorded before and after a 5 min pe-

riod of ischemia. RHBF, reactive hy-peremia blood flow. Total RHBF re-

fers to the shaded area under the curve.

Peak RHBF refers to the highestblood flow recorded after release ofarterial occlusion. C and D, blood flowrecorded before and after completionof 30 sec of rhythmic forearm exer-

cise. E and F, blood flow recordedbefore, during, and after 5 min of di-rect heating of the forearm.

962 R. Zelis, D. T. Mason, and E. Braunwald

a)

00

E

cm

-

o

-JLA-010

co

AL

401 B

(10) 3.

FORMAL a

wa]

Is> Cwo

I--(8 a

''CHF

'.01I c

10MINUTES OF ARTERIAL

(1

p<.OI <.01 <.01

1 3 5OCCLUSION

FIGURE 2 A, total reactive hyperemia blood flow (± SEM). B, peak reactive hy-peremia blood flow after release of 1, 3, 5, and 10 min of arterial occlusion. Solid

/ Circles, normal subjects. Open Circles, patients with congestive heart failure.

arm vascular bed to dilate in response to anischemic stimulus in the normal subjects and pa-tients with heart failure was compared on thebasis of two measurements: the total RHBF re-

sponse (Figs. 1A, 1B, and 2A) and the peakRHBF (Figs. 1A, 1B and 2B). The total RHBFin the patients with congestive heart failure wassubstantially less than in the normal subjects. The

60r-

I NI

50HLAJ

f->

U)

C,,-L 40c.01

Co

g E 3 0

a£cKI

5 E 20mE

__---cx

I-J-C)

:z0>

I 0

v CONTROL FOREARMVASCULAR RESISTANCE

N

p< .O1

N

N CHF

N.

15

4 a g 2,

_r-3

I -n

2 a

- -m r

= -j3."v-.:

MINIMAL FOREARMVASCULAR RESISTANCE

FIGuRE 3 Control and minimal forearm vascular resistance in normal subjects (closedcircles) and patients with congestive heart failure (open circles). Peak flow after releaseof 10 minm of arterial occlusion was used to calculate minimal forearm vascular resistance.

Vasodilator Stimuli in Heart Failure 963

A3J 3

L.c]a0

i0m-Jcl,,0w 2

I-

>-

0

I-

N

(II)'4y'A/

<.01 <.01

1 3 5

<.OI

10

N

NORMAL P<.O*

0F

ok

peak RHBFafter 10 min of ischemia approachesthe maximum that can be observed with an ische-mic stimulus of any duration. After this period ofischemia, the total RHBF in the patients withheart failure averaged only 31.9% of that observedin normal subjects, P < 0.01, (Fig. 2A). Simi-larly, the peak RHBF was markedly reduced inpatients with heart failure, as compared with nor-mal subjects. After 10 min of ischemia this valuein the patients with heart failure averaged 47.7%of that observed in the normal subjects (Fig. 2B).The minimal peripheral vascular resistance in theforearm estimated from the peak RHBF after10 min of ischemia is shown in Fig. 3. In normalsubjects peripheral vascular resistance fell to 2.06+ 0.14 mmHg per ml/min per 100 g, whereas inthe patients with heart failure the minimal re-sistance was significantly higher, and averaged4.05 + 0.56 mmHg per ml/min per 100 g (P< 0.01).

In order to determine whether the reducedRHBFof patients with heart failure resulted fromaugmented sympathetic constrictor activity, theeffects of nerve blockade on the response to

ischemia was studied in four patients with con-gestive heart failure. Peak RHBFwas essentiallyidentical before and after nerve blockade (Fig.4A); total RHBF was also not significantlyaffected. In four other patients with heart failure,adrenergic blockade with phentolamine similarlydid not alter peak RHBF (Fig. 4B) or totalRHBF in the calf. In four normal subjects theinfusion of norepinephrine raised the catechola-mine concentration of arterial blood to between1.39-2.44 ug/liter and elevated forearm resistancefrom 25.3 to 40.2 mmHg per ml/100 g per min.However, this heightened vasoconstrictor tone didnot interfere with the peak RHBF (Fig. 4C) orthe total RHBFresponse.

In order to determine whether the reducedRHBFin patients with heart failure was relatedto alterations in arteriolar tone produced by theelevated venous pressure, the effects of ischemiawere determined with and without prior venouscongestion. In seven normal subjects, total RHBFafter release of 10 min of arterial occlusion wasreduced by 9.1% with venous congestion, P <0.05, (Fig. 5A). In six patients with heart failure,

C H FNERVE BLOCK

FOREARM

AT

CONTROL ,i

-T

II- I~~~~~~~~~~~~~~~~~~1I

I NERVEBLOCK

p >. >.5 :.5I

3 5

CHF1.A. PHENTOLAMINE

CALF

60

8 so0-j

CO.w

< 3

w:w

a.Y 10.Ct

PHENTOL-AMINE

NORMALl.V. NOREPINEPHRINE

FOREARM

r_E - ~~~~~~TI

TI°TIEi

-

~I

0/T IOT

~9\CONTROL

p>.I,I

105

(5 min

MINUTES OF ARTERIAL OCCLUSION

FIGURE 4 A, peak reactive hyperemia blood flow after various periods of forearm ischemia before (solid circles)and after (open circles) deep nerve blockade of the forearm in patients with congestive heart failure. B, Peak re-

active hyperemia blood flow in the calf of patients with congestive heart failure after release of 5 min of arterialocclusion. White bar, control. Shaded bar, after intraarterial phentolamine. C, peak reactive hyperemia blood flowafter various periods of forearm ischemia in normal subjects before (solid circles) and during (open circles) an i.v.norepinephrine infusion.

964 R. Zelis, D. T. Mason, and E. Braunwald

20F

f15 _

1o0

3.0-JL

O00-JwooJ

4

w.

0l. -

a: o

XI sW E>- Eq

cr5

Ot,-, I

IVAL CIONGESLI A lLTE ITIA C-O4NLSION

00o 30--Jm

WcrW 0a.C ) 20-

W

w 10M

-JH

p <D5 6 ,>.2 60110 MINUTES VENOUS 10 MINUTES VENOUSARTERIAL CONGESTION ARTERIAL CONGESTIONOCCLUSION + OCCLUSION +

tO MINUTES 10 MINUTESARTERIAL OCCLUSION ARTERIAL OCCLUSION

FIGURE 5 Total reactive hyperemia blood flow after release of 10 min of ar-terial occlusion with and without 5 min of prior venous congestion in normalsubjects (panel A, solid bars) and in patients with congestive heart failure(panel B, shaded bars).

however, the total RHBF was not significantlyaltered by prior venous congestion (Fig. SB ).Furthermore, the total and peak RHBF afterischemia with venous congestion in normal sub-jects greatly exceeded the values observed inpatients with heart failure, regardless of whethervenous congestion had been carried out in them(Figs. 5A and 5B1).

III. Effects of compensation. The reactive hy-peremia response was restudied in five patientsafter compensation of their congestive heart fail-ure, an average of 17 days after the initial study.At the time of the second investigation these pa-tients all had normal venous pressures, a decreasein hepatomegaly, no edema, and diuresis hadresulted in weight loss that averaged 6.8 kg lessthan that present at the time of the first study.After compensation, total and peak RHBF weresignificantly higher following 1, 5, and 10 min ofarterial occlusion (Fig. 6).

IN". Postexercise hypereinia anid direct hleatinlg.The total forearm postexercise hyperemia bloodflow was markedly reduced in patients with heartfailure, as compared to normal subjects (Fig.7A), as was the total excess forearm blood flowindlucedl by direct heating (Fig. 7B).

V. Responise to exrogenious vasodilator stimuli.

Immediately after intra-arterial injection of phen-tolamine, blood flow rose to peaks which were sig-nificantly lower in patients with congestive heartfailure than in normal subjects (Fig. SB) ; andafter stabilization of blood flow following phentola-

15

0-J

000m

0

llJ tocr .!Wa-C.I0

0

E 5-

0

CHFl_

1-

"4-l

- I

IO4 >-CHF

p<.OI <.01

5

<.05

10MINUTES OF ARTERIAL OCCLUSION

FIGURE 6 The effect of compensation of heart failure onmodifying the total reactive hyperemia blood flow responseafter release of various periods of arterial occlusion.Open circles, heart failure-uncompensated. Solid circles,heart failure-compensated.

Vasodilator Stimuli in Heart Failure 965

FOREARMEXERCISE60

251

-J

as, 20

_j -Go 1-

(.) ->-&J 4

KI--X _

5

0

NORMAL CHF

FIGuRE 7 A, total postexercise hyperemia blood flow. This was calculated as the shaded areas underthe curves in Figs. iC and ID. B, total excess blood flow during 5 min of direct heating and 5 min ofrecovery. This was calculated as the shaded area under the curves in Figs. 1E and iF. Fine Strippledbars, normal subjects. Cross hatched bars, patients with congestive heart failure.

mine, it remained at a significantly lower level inthe patients with heart failure (Fig. 8C). The peakblood flow after intra-arterial injections of threedoses of NaNO2was significantly lower in patients

A B CCONTROL PEAK BLOOD BAS

BLOOD FLOW AFTER FLO

20 FLOW PHENTOLAMINE PHE

00E

-J

10 -

F3

with heart failure than in normal subjects (Fig.9), regardless of whether the NaNO2 was ad-ministered before or after phentolamine (Fig.8D).

N CHF N CHF N CHF N CHF N CHF

p<.OI p <.05 p<.OI p<.OI p<.OI

FIGURE 8 Peak calf blood flow response to intra-arterial vasodilators. A, control calfblood flow. B, peak calf blood flow after 1 mg of intra-arterial phentolamine. C, basalcalf blood flow after stabilization of blood flow after intra-arterial phentolamine. D,peak calf- blood flow response to 30 mg of intra-arterial sodium nitrite before (left)and after (right) adrenergic blockade with intra-arterial phentolamine. Fine stippledbars, normal subjects (N). Cross hatched bars, patients with congestive heart failure(CHF).

966 R. Zelis, D. T. Mason, and E. Braunwald

3:_C_

LL.

Ck_

V)

co

CL

>-I-xo

B

S.. ......

.. :..: .-......... . _: -::. .:......... .. : _

.,.: ... ........... ..... ..

.. ......_ :: : _... .:

........ ... , ..::. ...: .. .. : .

:::

::.::: :::::::::::::: :.... .... ........ ... :. _

:. -: _._ . 9 I,'WLDIRECT HEATING

1[

20

20 - I/

Normal

0%

- 5

10

C> /CHF T

co° 5UJ~~ ~ ~ ~~n6

L.1

0I10 30 50

SODIUM NITRITE DOSE (mg)

FIGURE 9 Peak calf blood flow response after intra-ar-terial injections of various doses of sodium nitrite. Solidcircles, normal subjects. Open circles, patients with con-gestive heart failure.

DISCUSSION

The major finding of this investigation is a strik-ing, heretofore unrecognized, reduction in the ca-pacity of the resistance vessels of patients withcongestive heart failure to respond to a widevariety of vasodilator stimuli. It was observed thatboth the peak and total RHBFfollowing variousperiods of ischemia were markedly reduced inpatients with congestive heart failure (Figs. 1A,1B, and 2). Consequently, after a maximal ische-mic stimulus, the calculated forearm vascular re-sistance in patients with heart failure declined tolevels approximately twice as high as those levelsobserved in normal subjects (Fig. 3). Also, thevasodilation, which normally occurs after muscularexercise and direct heating of the limb, was mark-edly attenuated by the heart failure state (Fig. 7).Exogenous humoral vasodilator stimuli, such asNaNO2and the immediate response to phentola-mine, an agent which not only blocks alphaadrenergic receptors but also exerts a direct, dilat-ing effect on smooth muscle (8), also evoked aless prominent vasodilating effect in patients withcongestive heart failure as compared with normalsubjects (Figs. 8 and 9). These differences oc-

curred despite the fact that the lower levels ofblood flow in the control state in patients withheart failure afforded a greater opportunity forexposure of the vascular smooth muscle to thesevasodilating drugs.

It is well known that forearm and calf bloodflow are reduced and the resistance in these vas-cular beds elevated in patients with congestiveheart failure in the basal state (1, 9, 10) and theresults of the present investigation are consistentwith these findings. Since, in the past, the highervascular resistance has been attributed to highersympathetic activity in patients with heart failure(11-14), we considered the possibility that thereduced responsiveness to dilator stimuli mightmerely be another manifestation of their increasedsympathetic vasoconstrictor activity. Indeed, re-cent reports have indicated that elevated sympa-hetic tone will decrease the reactive hyperemiaresponse (15, 16). In order to test this hypothesis,three types of experiments were undertaken (Fig.4). In the first two, sympathetic vasoconstrictortone to the limb was abolished in patients withcongestive heart failure by direct nerve blockadeand markedly reduced by the intra-arterial admin-istration of the alpha adrenergic blocking drugphentolamine. The effectiveness of the nerve block-ade was documented by observing that the vaso-constrictor response to the cold pressor test wasabolished and that, in fact, reversal occurred withforearm blood flow increasing with immersion ofthe opposite hand in ice water and control bloodflow tripling. The effectiveness of phentolaminewas reflected in the significant reduction of theconstrictor response to injected norepinephrineand to the cold pressor test. However, thesemaneuvers did not restore the response to dilatorstimuli to normal; the total and peak RHBFafteran ischemic stimulus were no higher after sympa-thetic blockade than before (Figs. 4A and B). It isunlikely that any residual sympathetic activityafter either method of blockade could have beenresponsible for the striking difference observedbetween the normal subjects and the patients withheart failure, in view of the finding of Strandelland Shepherd that peak active hyperemia bloodflow is little affected by intense reflex vasoconstric-tion (17). Furthermore, this postulation is con-sonant with the findings in experimental animalsby Remensnyder, Mitchell, and Sarnoff (18) that

Vasodilator- Stimuli in Heart Failure 967

even marked sympathetic vasoconstriction doesnot reduce vasodilatation resulting from muscularactivity.

In the third experiment, norepinephrine was in-fused into normal subjects and the circulatingcatecholamine concentration elevated to levels com-parable with those observed in patients with con-gestive heart failure during exhaustive exercise(12). Despite the fact that the maneuver elevatedforearm vascular resistance to levels observed inthe basal state in patients in heart failure, theRHBFwas not affected (Fig. 4C). From thesestudies it is concluded that sympathetically medi-ated vasoconstriction is not responsible for themarkedly abnormal vasodilator responses observedin heart failure.

A second possibility that must be considered toaccount for the difference between normal sub-jects and patients with heart failure is the presenceof occlusive vascular disease in the latter group,a significantly older population than the normalsubjects. Thus, if the lumen of the major arterieswere severely compromised, limb blood flow couldnot rise to normal levels after vasodilator stimulieven if the peripheral vessels exhibited a normaldilator response (19, 20). However, this possi-bility can also be excluded, since none of the pa-tients with heart failure in this study had ischemicor hypertensive heart disease, intermittent claudi-cation, or any manifestations of peripheral arterio-sclerosis. Finally, the diminished hyperemia re-sponses were observed in all extremities and wereas striking in the forearm as in the calf. The dif-ferences in their responses to vasodilator stimulicannot be attributed simply to these differences inage, since in both groups of individuals there wasno correlation between age and resting blood flow,peak and total RHBF, and peak response toNaNO2. All of the patients with heart failure werereceiving maintenance doses of cardiac glycosidesat the time of study. Since these drugs are knownto be mild vasoconstrictor agents (21) it was con-sidered that their administration might be respon-sible for a reduction in the vasodilator responses.However, this possibility was excluded by studiesin five additional subjects without heart failurewho received cardiac glycosides, but who exhibitednormal forearm vascular dynamics in control stateand after being exposed to vasodilator stimuli.

A fourth possibility considered to account for

the difference to the dilating stimuli relates to thedifference in venous pressure in the patients withheart failure and in normal subjects. The myogenichypothesis, proposed by Bayliss, suggests thatwhen intraluminal pressure falls after arterial oc-clusion, arteriole tone diminishes (22). It might beargued that the elevated venous pressure whichexists in patients with heart failure would preventintraluminal pressure from declining to the samelevel during occlusion in these patients, thus re-sulting in a lesser tendency for the arterioles todilate during the period of occlusion. That changein venous pressure can, in fact, alter the level ofblood flow after a period of ischemia was demon-strated in normal subjects (Fig. 5A) in whomRHBFwas slightly, though significantly, reducedwhen the veins were congested before occlusionof arterial inflow; and this response is consistentwith that described by Wood et al. (6). However,this reduction in RHBFby elevating venous pres-sure markedly was very slight and did not lowerRHBFto those levels observed in the presence ofheart failure (Fig. 5). Furthermore, when a com-parison was made between the RHBF in normalsubjects and patients with heart failure, in whomvenous pressure had been elevated to similarlevels, the aforementioned differences persisted.Therefore, whereas the elevated venous pressureof heart failure may play a small role in the re-duced RHBF, it certainly cannot be consideredresponsible for the major changes observed. Also,elevation of venous pressure would not explainthe diminished response exhibited by these patientsto the several other vasodilator stimuli tested.

The possibility was also considered that reduc-tion in the diffusion of vasodilator metabolites tothe vascular wall might account for the impairmentin the vasodilator response observed after ischemiaand exercise (25). Certainly, if the interstitialfluid volume were greatly augmented this couldoccur. However, the finding that vasodilation wasas reduced in the arms as in the legs would militateagainst this possibility. More importantly, the find-ing that vasodilation was reduced after the intra-arterial administration of two exogenous vaso-dilators, sodium nitrite and phentolamine, pointsto an abnormality in the vessel wall itself. Indeed,the lower blood flow existing in patients withheart failure insured that their resistance vesselswere actually exposed to these vasodilators in

968 R. Zelis, D. T. Mason, and E. Braunwald

higher concentrations and for longer periods oftime than normal subjects.

In considering the last named possibility, i.e.that the difference in vascular responsiveness re-sults from a structural abnormality in the wall ofthe resistance vessels in heart failure, studies inhypertensive subjects must be considered. Tobian,Janecek, Tomboulian, and Ferreira and Olsonand Chesley have demonstrated an increased so-dium and water content in the arterioles of hyper-tensive rats and humans, and suggested that thismight produce a structural abnormality in thevessel and increase resistance to flow (24, 25).Conway's finding of a reduced reactive hyperemiaresponse in the forearm of patients with hyper-tension is consistent with this suggestion (26).It is possible that in heart failure the resistancevessels participate in the generalized retention ofsalt and water. This could be accompanied byswelling, engorgement, and stiffening of thesevessels, which, in turn, could lead to an inabilityto dilate appropriately when stimulated. The ob-servation that treatment of hear failure by diuresistended to restore the dilating response toward nor-mal (Fig. 6) provides support for this suggestion.In any event, further studies particularly in ani-mals with experimentally produced heart failure,will be required to throw further light on thisquestion.

Regardless of the specific mechanism responsiblefor the marked reduction in the response to vaso-dilator stimuli which exists in patients with con-gestive heart failure, it is interesting to speculateupon the possible physiologic effects of this abnor-mality. In heart failure the dilatation of vascularbeds supplying skeletal muscle would tend to bereduced during and after the contraction of thesemuscles. As a consequence, arterial pressure mightnot fall despite an inadequate elevation of cardiacoutput. A larger fraction of the cardiac outputwould then be available for perfusion of morevital organs than would be the case if the resist-ance vessels supplying the exercising muscles di-lated normally. Presumably, increased extraction ofoxygen by the peripheral tissues could compensatefor the lack of dilation that occurs in patients withheart failure (10). This would be more economicalin the face of a limited total cardiac output thanthat of the normal dilatation, which may not beessential (27).

REFERENCES1. Hewlett, A. W., and J. G. Van Zwaluwenburg. 1909.

The rate of blood flow in the arm. Heart. 1: 87.2. Whitney, R. J. 1953. The measurement of volume

changes in human limbs. J. Physiol. (London). 121: 1.3. Holling, H. E., H. C. Boland, and E. Russ. 1961.

Investigation of arterial obstruction using a mercury-in-rubber strain gauge. Am. Heart J. 62: 194.

4. Kerslake, D. McK. 1949. The effect of the applicationof an arterial occlusion cuff to the wrist on the bloodflow in the human forearm. J. Physiol. (London).108: 451.

5. Euler, U. S. von, and F. Lishajko. 1961. Improvedtechnique for the fluorometric estimation of catechola-mines. Acta Physiol. Scandinav. 51: 348.

6. Wood, J. E., J. Litter, and R. W. Wilkins. 1955.The mechanism of limb segment reactive hyperemiain man. Circulation Res. 3: 581.

7. Bodansky, 0. 1951. Methemoglobinemia and methemo-globin-producing compounds. Pharmacol. Rev. 3: 144.

8. Taylor, S. H., G. R. Sutherland, G. J. MacKenzie,H. P. Staunton, and K. W. Donald. 1965. The cir-culatory effects of intravenous phentolamine in man.Circulation. 31: 741.

9. Braunwald, E., C. A. Chidsey, P. E. Pool, E. H.Sonnenblick, J. Ross, Jr., D. T. Mason, J. F. Spann,Jr., and J. W. Covell. 1966 Congestive heart failure.Biochemical and physiological considerations. Ann.Internal Med. 64: 904.

10. Wade, 0. L., and J. M. Bishop. 1962. Cardiac Outputand Regional Blood Flow. Blackwell Scientific Pub-

lications, Ltd., Oxford. 134.11. Starling, E. H. 1897. Points on pathology of heart

disease. Lancet. 1: 569.12. Donald, K. W., J. M. Bishop, and 0. L. Wade. 1955.

Changes in the oxygen content of axillary venousblood during leg exercise in patients with rheumaticheart disease. Clin. Sci. 14: 531.

13. Gaffney, T. E., and E. Braunwald. 1963. Importanceof the adrenergic nervous system in the support ofcirculatory function in patients with congestive heartfailure. Am. J. Med. 34: 320.

14. Chidsey, C. A., E. Braunwald, and A. G. Morrow.1965. Catecholamine excretion and cardiac stores ofnorepinephrine in congestive heart failure. Am. J.Med. 39: 442.

15. Abrams, M. E., D. J. P. Barker, and W. J. H.Butterfield. 1965. The effect of reserpine, noradrena-line and adrenaline on reactive hyperemia in thehuman forearm. Clin. Sci. 29: 565.

16. Ardill, B. L., V. M. Bhatnagar, and P. H. Fentem.1966. The suppression of reactive hyperemia by theaction of sympathetic adrenergic nerves. J. Physiol.(London). 187: 24 P. (Abstr.)

17. Strandell, T., and J. T. Shepherd. 1967. The effect inhumans of increased sympathetic activity on the bloodflow to active muscle. Acta. Med. Scand. Suppl. 472:146.

Vasodilator Stimuli in Heart Failure 969

18. Remensnyder, J. P., J. H. Mitchell, and S. J. Sarnoff.1962. Functional sympatholysis during muscular ac-tivity. Circulation Res. 11: 370.

19. Shepherd, J. T. 1960. The blood flow through thecalf after exercise in subj ects with arteriosclerosisand claudication. Clin. Sci. 9: 49.

20. Winsor, T. 1951. Simplified determination of arterialinsufficiency. Circulation. 3: 830.

21. Mason, D. T., and E. Braunwald. 1964. Studies ondigitalis X. Effects of ouabain on forearm vascularresistance and venous tone in normal subjects and inpatients in heart failure. J. Clin. Invest. 43: 532.

22. Bayliss, W. M. 1902. On the local reactions of thearterial wall to changes of internal pressure. Physi-ology 28: 220.

23. Paterson, G. C., and R. F. Whelan. 1955. Reactivehyperaemia in the human forearm. Clin. Sci. 14: 197.

24. Tobian, L., J. Janecek, A. Tomboulian, and D. Fer-reira. 1961. Sodium and potassium in the walls ofarterioles in experimental renal hypertension. J. Cin.Invest. 40: 1922.

25. Tobian, L., R. Olson, and G. Chesley. 1967. Watercontent of arteriolar wall in renal hypertension. Fed-eration Proc. 26: 377. (Abstr.)

26. Conway, J. 1963. A vascular abnormality in hyper-tension. A study of blood flow in the forearm. Circu-lation. 27: 520.

27. Blair, D. A., W. E. Glover, and I. C. Roddie. 1959.The abolition of reactive and post-exercise hyperemiain the forearm by temporary restriction of arterialinflow. J. Physiol. (London). 148: 648.

970 R. Zelis, D. T. Mason, and E. Braunwald