The Dicrotic Arterial PulseBy GORDON A. Ewy, M.D., JORGE C. Rios, M.D., AND FRANK I. MARCUS, M.D.

SUMMARYThe study consisted of observations on nine male patients with palpable dicrotic

carotid pulses. Average patient age was 37 years. All had advanced myocardial failureas evidenced by cardiomegaly and the presence of prominent atrial and ventriculardiastolic gallop sounds. Most of these patients were diagnosed as having primarymyocardial disease.The indirect carotid pulse was characterized by a single systolic wave, a low dicrotic

notch, and a large dicrotic wave. The direct brachial arterial pressure pulse had asimilar configuration with a shortened ejection time index. The hemodynamic data onthese patients was characterized by low cardiac output, low stroke volume, elevatedpulmonary arterial wedge pressures, and high total systemic resistance. From theseobservations we conclude that the presence of a marked dicrotic pulse in afebrilepatients at rest may indicate severe functional impairment of the myocardium.

Additional Indexing Words:Dicrotic notch Dicrotic wavePrimary myocardial disease Hemodynan

THE dicrotic arterial pulse is character-ized by two pulsations with each cardiac

cycle; the second is due to an accentuateddicrotic wave (fig. 1). Paul Wood wrote,". . . the dicrotic pulse is encountered chieflyin patients sick with fever such as typhoid; theperipheral resistance is low, the arteries arelax, and the cardiac output probably normal."'During the past 3 years we have observed anumber of patients with a palpable dicroticarterial pulse. None of them were febrile. Itwas our clinical impression that these patientshad severe impairment of myocardial function.Accordingly, nine of these patients wereassessed.

From the Georgetown University School ofMedicine, Department of Medicine, Division ofCardiology, District of Columbia General Hospital,and the George Washington University School ofMedicine, George Washington University MedicalDivision, District of Columbia General Hospital,Washington, D. C.

This study was supported in part by grants fromthe National Heart Institute, National Institutes ofHealth, HE-5003 and HE-5433 and from the SpecialCardiac Research Fund, Georgetown University. Dr.Marcus is the recipient of U. S. Public Health CareerDevelopment Award HE-25,575.Circulation, Volume XXXIX, May 1969

nic dataPeripheral vascular resistance

MethodsEach of the nine male patients studied had a

dicrotic carotid pulse that was readily palpable.They ranged in age from 29 to 48 years with anaverage age of 37 years. Seven had primarymyocardial disease, one (Z.C.) had advancedhypertensive cardiovascular disease, and one(C.R.) had coronary artery disease with a largeventricular aneurysm. Two of the patients (Z.C.and H.M.) had associated pulmonary emboli. Allpatients had cardiomegaly by physical androentgenographic examination. All had both atrialand ventricular, or summation diastolic gallopsounds on auscultation and phonocardiograms;none had cardiac murmurs. All were in normalsinus rhythm.

Controls for the indirect carotid pulse record-ings were obtained from nine normal malesubjects who ranged in age from 25 to 38 yearswith an average age of 29 years. Controls for thehemodynamic measurements and direct brachialarterial tracings were obtained from 15 normalsubjects previously catheterized and reported onby Perloff and associates.2 They ranged in agefrom 16 to 36 years with an average age of 25years. Differences were analyzed by the two-tailed Student t-test.3

Indirect carotid arterial pulse recordings weremade in seven of these patients utilizing an air-filled funnel 2.5 cm in diameter that wasconnected to a Statham physiologic transducer(model PR 23-2G-300) by a 10-cm polyethylene

655

by guest on June 7, 2018http://circ.ahajournals.org/

Dow

nloaded from

EWY ET AL.

Phono S1 S2 S3.. ,; . . .. A A

*

,DG

Indi rectCarotid

P QRS

R.L.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ECG TT

R . L.

Figure 1

Indirectly recorded tracing of the carotid artery pulseshowing dicrotic arterial pulse. Note that one im-pulse is in systole and the second is in diastole. Thedicrotic pulse should be differentiated from the bis-feriens arterial pulse in which both palpable impulsesare systolic. From top to bottom: low-frequencyapical phonocardiogram, indirect carotid arterial pulsetracing, and electrocardiogram.

tubing. Recordings were made on an Electronicsfor Medicine DR 8 recorder at paper speeds of 50mm/sec or greater. Since the level of the dicroticnotch is influenced by the time-constant of therecording instrument, it is important to note thatthe recording system used has a time-constantapproaching infinity.

Three measurements were made on both theindirect carotid and direct brachial arterialpressure pulse recordings in an effort to quanti-tate the differences in wave form that are obviouswhen comparing the dicrotic and the normalpulse (fig. 2). The first was to measure the peakheight of the dicrotic wave (d) above the level ofthe dicrotic notch and to express this height as a

percentage of the total height of the systolic pulse(t) measured from the onset or foot of the pulseto the peak of the systolic wave (fig. 2). Thesecond was to measure the level of the dicroticnotch (n) above the onset or foot of the systolicwave and to express this distance as a percentageof the total height of the systolic pulse (t). Thethird was to express the area of the dicrotic waveabove the level of the dicrotic notch (fig. 2, blackarea) as a percentage of the area of the systolic

t DICROTIC

2I

3 4

=.-:J- t NOMAL

"N. _ _

Figure 2Direct brachial arterial pressure pulse recordings. Ascompared to the normal, the dicrotic pulse, as illus-trated in panels 1 and 2, is characterized by a lowlevel of the dicrotic notch (n) and an increased heightof the dicrotic wave (d). The area under the dicroticwave (black) relative to the area under the systolicwave (dashed lines) is larger in the dicrotic than inthe normal pressure pulse, iUustrated in panels 3 and 4.

wave (fig. 2, dashed area). Since the duration ofejection appeared short, the ejection time indexwas also measured. The ejection time index is thesystolic ejection time corrected for heart rate bythe method of Weissler and associates.4

Right heart catheterization was performed inthe standard manner, and pressures were ob-tained with the use of a Statham 23 Dbphysiologic transducer and an Electronics forMedicine DR 8 recorder. Left heart catheteriza-tion and left ventricular angiograms were ob-tained in the one patient (C. R.) who had aventricular aneurysm. Cardiac output was de-termined in duplicate by either the Fick principleor by the dye-dilution technic, injecting indo-cyanine-green indicator into the pulmonaryartery and sampling from either the brachial orfemoral artery.The effect of altering the arterial resistance on

the dicrotic pulse was studied in three patients bydistal occlusion of the artery. The brachial orfemoral artery was occluded distal to the recordingsite by placing a blood pressure cuff around theforearm or thigh and inflating it to a pressureabove systemic arterial pressure. In one of thesethree patients, angiotensin was injected throughthe same Cournand needle from which the pres-sure pulse was recorded. In another, a vasodilator,tolazoline, was infused in a similar manner. Inaddition, the effect of amyl nitrite inhalation wasrecorded in three patients.

Circulation, Volume XXXIX, May 1969

IL ..~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~----6--W-000-Mom -4..10 1-MAP . p A A 10 a 1 A 'W--I

656

by guest on June 7, 2018http://circ.ahajournals.org/

Dow

nloaded from

DICROTIC ARTERIAL PULSE

IIII ILrrn I III

mEellr XX 11 r T Tr r

i T §Tl |lf

Ial-1I11Tg

NoXlXrmalT

C___ _ _ _ _ _ _

_ _ _ _ _ _ _ __ _ _ _ _ _ _ _

_ _ _ _ _ _ _ __ _ _ _ _ _ _ J_ _ _ _ _ _ z r_ _ _ _ _ _ t __ _ _ _ _ _ r __ _ _ _ _ ] _ __ _ _ _ I _ __ _ _ _ _ C _ __ _ n _ _

i | | _ f _ _ __ _ _ _ _ __ _ _ _ _ _ _ __ _ _ _ _ _ _ __ _ _ _ _ _ _ _

_ _ _ _ _ _ __ _ _ _ _ _ __ _ _ _ _ _ _

_ _ _ _ _ _ _

D I17 . L

NormalFigure 3

The indirectly recorded carotid pulse. These record-ings, two from patients (above) and two from normals(below) were made utilizing an air-filled funnelcoupled to a Statham physiologic strain-gauge trans-ducer. Note that the dicrotic pulses are characterizedby a single systolic wave, a low dicrotic notch, anda large dicrotic wave. The configuration of the normalindirectly recorded carotid pulse is illustrated belowfor comparison.

Results

Indirect Carotid Pulse

Two indirect recordings of the carotidarterial pulse from these patients are contras-ted with two normal tracings in figure 3. Thecontour of the systolic wave of the indirectcarotid pulse of these patients was character-ized by a single peak. This configurationcontrasts with the two maxima found in theindirect pulse recordings from normal sub-jects.5 The level of the dicrotic notch in thepatients was low, averaging 9% of the totalCirculation, Volume XXXIX, May 1969

pulse height (n/t) in contrast to the averagelevel of 46% in the normal. The dicrotic wavewas prominent. The height of the dicroticwave averaged 49% of the total pulse height(d/t) in these patients compared to 6% in thenormal. The area of the dicrotic waveaveraged 44% of the area of the systolic wave inthe patients in contrast to 2% in normals.

Direct Brachial Arterial Pulse

The brachial arterial pressure pulse wasrecorded directly in eight of the nine patients.The level of the dicrotic notch was low,averaging 10% of the total height of therecorded pulse compared to the average levelof 40% in the normals (fig. 2). The height of thedicrotic wave was increased, averaging 44% ofthe total height of the pressure pulse incontrast to an average height of 4% in thenormals. As in the indirect carotid pulse, thearea of the dicrotic wave was increasedcompared to the area of the systolic wave inthe directly recorded brachial arterial pressurepulse; 46% in these patients and 1.3% in thenormals. The average ejection time index was0.04 sec shorter than the normal mean value of0.42 sec of the controls (P = < 0.001).

Distal occlusion of the artery by a bloodpressure cuff or locally increasing the distalconstriction by intra-arterial infusion of angio-tensin resulted in an accentuation of thedicrotic wave (fig. 4). Inhalation of amylnitrite resulted in a decrease in the height ofthe dicrotic wave and either no change or anelevation of the level of the dicrotic notch (fig.4).

Cardiac Catheterization

The hemodynamic findings are tabulated intable 1. The cardiac output was low, with amean cardiac index of 1.74 L/min/m2. Thestroke volume was small, with a mean strokeindex of 17 cc/beat/im2. The pulmonary arteri-al wedge or the left ventricular end-diastolicpressures or both, were elevated with a meanpressure of 24.5 mm Hg. The mean total sys-temic resistance of 2,152 dyne-sec cm-'' wastwice as high as the mean value for the con-trol patients.

II

A

IIIIIIIIIIII

.IIIIIIIIIIIIIIIII

IIIIIIIIIIIIIIIIIIIII

..II

.IIIIIIIIIIIIII

tttt11:1Ziiii

I

___:I

I

-

II

!

657

.1

II

i

I-TTT-MI

R.B

ML

by guest on June 7, 2018http://circ.ahajournals.org/

Dow

nloaded from

EWY ET AL.

00o

80

IBA BA

I ~~~~~~~~~~~~~I I BA I

0

BA

FA

CONTROL

L.L.

FA

ANGIOTENSIN( by brochial ortery infusion)

FA

AMYL NITRITE(inhalation)

100

_ 80

60

40

20

FA

Figure 4

From top to bottom: electrocardiogram, direct brachial arter pulse (BA) with pressure cai-brations on the left, and direct femoral arterial pulse (FA) with pressure cdations on theright. Note that the FA pressure pulse has a smauler dicrotic wave than the simultaneouslyrecorded BA. (Left panel) Control recordings of BA and FA pressure pulses. (Middle panel)The effect of infusion of angiotensin into BA to increase arterial resistance. Note the increasein the height of the dicrotic wave of the BA tracing and the lowered dicrotic notch. (Rightpanel) BA and FA recordings following inhalation of amyl nitrite. The dicrotic wave is smaller.

Table 1

Hemodynamic Data from Nine Patients with Dicrotic Arterial Pulses

Age BSA Mean BP (mm Hg) CI SI TSR

Patient (yr) Diagnosis (m 2) BA RA Wedge (L/min/m2) (cc/beat/m2) dyne-sec cm -5

Z.C. 33 HCVD 2.20 120 7 28 1.68 20 2,591R.L. 32 PMD 1.38 73 11 - 1.91 15 2,209J.W. 33 PMD 1.64 75 15 18 2.24 20 1,628R.B. 35 PMD 1.85 84 8 25 2.54 26 1,423L.E. 29 PMD 2.15 80 20 32 1.18 12 2,507C.R. 46 CAD-VA 1.68 68 27 1.19 12 2,717L.L. 38 PMD 1.97 83 12 27 1.77 19 1,906E.H. 38 PMD 2.02 87 11 24 1.85 15 1,854H.M. 48 PMD 2.15 89 11 15 1.32 13 2,530

Average:Patients 37 1.89 84 12 24.5 1.74 17 2,152SD ±0.28 + 15 i4.1 45.5 ±0.46 i 4.7 i466Normals 25 1.79 88 5 12 3.79 51 1,099SD ±0.32 ±13 ±t0.85 ±13 =±404P 0.5 N.S. 0.001 0.001 0.001

Abbreviations: BSA = body surface area; m = meter; BP = blood pressure; BA = brachial artery; RA = rightatrium; wedge = pulmonary arterial wedge; CI = cardiac index; SI = stroke index; TSR = total systemic resistance;HCVD = hypertensive cardiovascular disease; PMD = primary myocardial disease; CAD-VA = coronary arterydisease with ventricular aneurysm.Normal values were calculated from the data obtained by Perloff and associates from 15 subjects.2 RA and wedge

pressure measurements were not obtained from the control subjects. The values listed represent the upper limits ofnormal for our laboratory.

Cirecuaion, Volume XXX1X, Moy 1969

658

I- r

by guest on June 7, 2018http://circ.ahajournals.org/

Dow

nloaded from

DICROTIC ARTERIAL PULSE

DiscussionThe English literature has few references to

the dicrotic arterial pulse. Textbooks ofcardiology" 6 refer to the work of Sir JamesMacKenzie. In 1902, MacKenzie published hisclassic book, The Study of the Pulse, Arterial,Venous and Hepatic and of the Movement ofthe Heart, in which he wrote: "When theventricular contraction is weak the systolicwave is low, the dicrotic notch is deep,reaching often to the base line, while thedicrotic wave is much increased in size,indicating a remarkable fall in arterial pres-

sure." He illustrated the dicrotic pulse in a

patient with typhoid fever and attributed themarked increase in the dicrotic wave to thefact that".. the arterial relaxation was muchgreater in the typhoid case."7 Paul Wood alsofound this type of arterial pulsation in patientswith infectious or toxic reactions. He wrote,".9. .a good sign of vascular relaxation is a

markedly dicrotic pulse and although notnecessarily serious should put the physician on

guard."'Our studies indicate that the dicrotic pulse

occurs in young patients who have elevatedperipheral vascular resistance and a lowstroke volume. Although these observationsappear to contrast with those in the litera-ture,"1 6, 7 it is tempting to speculate that thoseearlier observations were made on patients

E C G

B A .

Figure 5

Electrocardiogram (E.C.G.) and direct recording ofthe brachial arterial pulse (B.A.). Note that the post-premature beat has a more rapid upstroke velocity,a larger systolic wave, a higher dicrotic notch, anda smaller dicrotic wave than the control beat occurringbefore the extrasystole.

Circulation, Volume XXXIX, May 1969

who had either typhoid myocarditis or a lowstroke volume secondary to severe dehydra-tion from typhoid diarrhea.The determinants of the dicrotic wave are

not completely understood. The dicrotic waveof the normal arterial pressure pulse has beenascribed to the rebound of arterial bloodagainst a closed aortic valve. However,peripheral factors contribute materially to itsformation.8-12 There is a decrease or loss of thedicrotic wave with age, hypertension, general-ized atherosclerosis, and diabetes.5 10 13 Ourpatients were relatively young, with a meanage of 37 years. It may be that the arterialresiliency of the young is necessary for thegeneration of the dicrotic pulse.What other factors contribute to the forma-

tion of the dicrotic pulse? Increasing localarterial resistance in our patients heightenedthe dicrotic wave. In the one patient studied,local injection of an arterial dilator resulted ina decrease of the dicrotic wave. Theseobservations confirm the importance of pe-ripheral factors in the genesis of the dicroticwave. Three other observations bear on thegenesis of the dicrotic pulse: The first is thatthe post-premature beats were less dicroticthan the control beats (fig. 5). In the absenceof muscular subaortic stenosis, the moreforceful contraction of the post-prematurebeat results in an increased stroke volume.Second, in our patients with pulsus alternans,the alternate beats with a higher pressure hada smaller dicrotic wave and a higher dicroticnotch than the beats with lower pressure. Thethird observation was in a patient who hadpericardial tamponade and a dicrotic arterialpulse only on inspiration (fig. 6A). Intamponade it is known that the stroke volumedecreases with inspiration.14 In addition, theejection time index (an indirect measurementof stroke volume) of the dicrotic beats was0.04 sec shorter than that of the more normalbeats (fig. 6B). The dicrotic contour of thebrachial arterial pulse was not present follow-ing the removal of 500 ml of pericardial fluid.These observations seem to indicate that asmall stroke volume or a weaker force of

659

by guest on June 7, 2018http://circ.ahajournals.org/

Dow

nloaded from

EWY ET AL.

f40

'PlT -0 l

'P

A-Fi BFigure 6

Effect of respiration on brachial arterial pulse contour in a uremic patient with pericardialtamponade.

(A) During tamponade. From top to bottom: electrocardiogram, direct brachial arterialpressure pulse (BA), intrapericardial pressure recording (IP) and respiration: inspiration (in),giving a downward deflection, and expiration (ex), giving an upward deflection. Note themarked dicrotic pulse (arrows) on inspiration, when stroke volume would be small and thealmost normal contour of the pulse during expiration when the stroke volume would be larger.The intrapericardial pressure is 20 mm Hg.

(B) Following removal of 500 ml of pericardial fluid by technic previously reported.15From top to bottom: electrocardiogram, direct recording of brachial arterial pressure pulse(BA), respiration, and intrapericardial pressure (IP). Following the relief of tamponade, notethe absence of the dicrotic pulse even during inspiration. The intrapericardial pressure de-creased to a mean of 5 mm Hg with respiratory fluctuations between 11 and -1 mm Hg.

contraction, or both, may also be important inthe formation of the dicrotic pulse.These observations concerning the role of

the arterial resiliency, peripheral resistance,stroke volume, or force of myocardial contrac-tion in the genesis of the dicrotic arterial pulseare all in accord with the findings in our ninepatients. Those patients were relatively young,with elevated peripheral resistance and lowstroke volumes. All had clinical and hemody-namic evidence of impairment of myocardialfunction.

The dicrotic arterial pulse appears to bedetermined by the presence of certain hemo-dynamic factors in relatively young patients.Therefore, the dicrotic pulse is not necessarilyrelated to a specific disease entity as reportedby Meadows and co-workers.'6 We haveobserved the dicrotic arterial pulse in a varietyof disease states, including primary myocardi-al disease, hypertensive cardiovascular dis-ease, ischemic heart disease, acute myocardialinfarction, primary pulmonary hypertension,and intermittently in a patient with pericardi-

Circulation, Volume XXXIX, May 1969

I|--@lw

I ll

660O

by guest on June 7, 2018http://circ.ahajournals.org/

Dow

nloaded from

DICROTIC ARTERIAL PULSE

al tamponade. A dicrotic arterial pulse hasbeen reported in constrictive pericarditis.17

In view of these studies, whenever adicrotic pulse is palpated in an afebrilepatient, we are oriented to think of severefunctional impairment of the myocardium.

AcknowledgmentWe are indebted to Dr. A. C. deLeon for making

the hemodynamic data on the normal patientsreported by Perloff and associates2 available foranalysis. Dr. Nayab Ali made the recording used infigure 6. Drs. A. C. deLeon, J. K. Perloff, JamesRonan, and R. A. Massumi kindly reviewed themanuscript. The authors appreciate the help of Dr.George Just of Mainz, Germany, for referring us tothe work of Dr. Kurt Wezler and to Volker Sonntagfor the German translations. The secretarial assistanceof Mrs. Yvonne McIlwain is gratefully appreciated.

References

1. WOOD, P.: Diseases of the Heart and Circulation,ed. 2. Philadelphia, J. B. Lippincott Co., 1956,pp. 28, 610.

2. PERLOFF, J. K., CALVIN, J., DE LEON, A. C., ANDBOWEN, P.: Systemic hemodynamic effects ofamyl nitrite in normal man. Amer Heart J 66:460, 1963.

3. YOUDEN, W. J.: Statistical Methods for Chemists.New York, John Wiley & Sons, 1951, pp. 24,119.

4. WEISSLER, A. M., HARRIs, L. C., AND WHITE, G.D.: Left ventricular ejection time index inman. J Appl Physiol 18: 919, 1963.

5. FREIs, E. D., HEATH, W. C., LUCHSINGER, P. C.,AND SNELL, R. E.: Changes in the carotid pulsewhich occur with age and hypertension. AmerHeart J 71: 757, 1966.

6. HURST, J. W., AND LOGUE, R. B.: The Heart,

Arteries and Veins. New York, McGraw-HillBook Co., 1966, p. 73.

7. MAcKENzIE, J.: The Study of the Pulse, Arterial,Venous and Hepatic and of the Movement ofthe Heart. London, Young J. Pentland, 1902, p.25.

8. FEINBERG, A. W., AND LAX, H.: Studies of thearterial pulse wave. Circulation 18: 1125,1958.

9. HAMILTON, W. F.: Patterns of the arterialpressure pulse. Amer J Physiol 141: 235,1944.

10. HONIG, C. R., TENNEY, S. M., AND GABEL, P. V.:Mechanism of cardiovascular action of nitro-glycerine. Amer J Med 29: 910, 1960.

11. ALEXANDER, R. S.: Factors determining thecontour of the pressure pulses recorded fromthe aorta. Fed Proc 2: 738, 1952.

12. BECK, W., SCHRIRE, V., VOGELPOEL, L., NELLEN,M., AND SWANEPOEL, A.: Hemodynamic effectsof amyl nitrite and phenylephrine on normalhuman circulation and their relation to changesin cardiac murmurs. Amer J Cardiol 8: 341,1961.

13. O'RouRKE, M. F., BLAZEK, J. V., MORREELS, C.L., JR., AND KROVETZ, L. J.: Pressure wavetransmission along the human aorta. Circula-tion Research 23: 567, 1968.

14. SHABETAI, R., FOWLER, N. O., AND GUERON, M.:Effects of respiration on aortic pressure andflow. Amer Heart J 65: 525, 1963.

15. MASsUMI, R. A., Rios, J. C., Ross, A. M., ANDEwY, G. A.: Technique for insertion of anindwelling intrapericardial catheter. Brit HeartJ 30: 333, 1968.

16. MEADOWS, W. R., OSADJAN, C. E., AND SHARP, J.T.: Dicrotic pulse of primary myocardialdisease. (Abstr.) Circulation 36 (suppl. II): II-184, 1967.

17. WEZLER, K., AND GREVEN, K.: Uber dieEnstehung des Dikroten (Resonanz-) Pulses. ZGes Exp Med 105: 540, 1939.

Circulation, Volume XXXIX, May 1969

661I

by guest on June 7, 2018http://circ.ahajournals.org/

Dow

nloaded from

AMERICAN HEART ASSOCIATION

Circulataon, Volume XXXIX, May 1969

662

by guest on June 7, 2018http://circ.ahajournals.org/

Dow

nloaded from

GORDON A. EWY, JORGE C. RIOS and FRANK I. MARCUSThe Dicrotic Arterial Pulse

Print ISSN: 0009-7322. Online ISSN: 1524-4539 Copyright © 1969 American Heart Association, Inc. All rights reserved.

is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Circulation doi: 10.1161/01.CIR.39.5.655

1969;39:655-662Circulation. 

http://circ.ahajournals.org/content/39/5/655located on the World Wide Web at:

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

  http://circ.ahajournals.org//subscriptions/

is online at: Circulation Information about subscribing to Subscriptions: 

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

  document. and Rights Question and Answer

Permissionsthe Web page under Services. Further information about this process is available in thewhich permission is being requested is located, click Request Permissions in the middle column ofClearance Center, not the Editorial Office. Once the online version of the published article for

can be obtained via RightsLink, a service of the CopyrightCirculationoriginally published in Requests for permissions to reproduce figures, tables, or portions of articlesPermissions:

by guest on June 7, 2018http://circ.ahajournals.org/

Dow

nloaded from