RItcwtarnwflEidp

Rttertt(tfetd

M

eatc

FHvbtGr

G1m

2

Effect of Tirofiban on PercutaneousCoronary Intervention-Induced

Endothelial Dysfunction in Patients WithStable Coronary Artery Disease

Ascan Warnholtz, MD, Mir Abolfazl Ostad, MD, Thomas Heitzer, MD,

Britta U. Goldmann, MD, Goetz Nowak, MD, and Thomas Munzel, MD

tfias0i4doFoetc

ecent studies demonstrated that glycoprotein (GP) IIb/IIa receptor antagonists improve endothelial dysfunc-ion of forearm resistance vessels in patients with stableoronary artery disease. However, it remains unclearhether these findings can be extended to the conduc-

ance vessel level. In this study, we aimed to evaluate thecute effect of tirofiban on endothelial function of arte-ial conductance vessels in patients undergoing percuta-eous coronary intervention (PCI). Endothelial functionas examined by ultrasonographic measurement ofow-mediated vasodilation (FMD) of the brachial artery.ndothelium-independent vasodilation was determinedn response to nitroglycerin. Sixty-six patients who un-erwent PCI were included in the study. Thirty-three

atients received a bolus of 10 �g/kg body weight of

dIh

ditLrslswofmIaaflvtwvmHttDAaoainz.de.

0 ©2005 by Excerpta Medica Inc. All rights reserved.The American Journal of Cardiology Vol. 95 January 1, 2005

irofiban, whereas 33 patients who did not receive tiro-ban served as the control group. FMD was measured inll patients before and 30 minutes after PCI. Tirofibanignificantly improved FMD (6.0 � 0.4% before vs 7.8 �.5% after PCI, p <0.0001), whereas FMD deteriorated

n patients in the control group (6.1 � 0.6% before vs.7 � 0.7% after PCI, p � 0.006). Nitroglycerin-inducedilation remained unaltered in response to PCI. In an-ther group of 11 patients with coronary artery disease,MD did not change after coronary angiography with-ut coronary intervention. In conclusion, PCI inducesndothelial dysfunction in forearm conductance vesselshat can be reversed with tirofiban. �2005 by Ex-erpta Medica Inc.

(Am J Cardiol 2005;95:20–23)

esults from recent studies suggest a beneficialeffect of glycoprotein (GP) IIb/IIIa receptor an-

agonists on endothelial function of the microcircula-ion in patients with coronary heart disease.1,2 How-ver, published reports on the effect of GP IIb/IIIaeceptor blockade on endothelial function in conduc-ance vessels are lacking. In addition, it is not knowno what extent percutaneous coronary interventionPCI) may influence endothelial function distant fromhe intervention site. The aim of this study was there-ore twofold: (1) to test whether PCI may inducendothelial dysfunction in the brachial artery, and (2)o examine the effect of tirofiban during PCI on en-othelial function of the brachial artery.

ETHODSPatient group: Sixty-six patients who underwent

lective PCI for symptomatic coronary artery diseaset the University Hospital Hamburg-Eppendorf be-ween December 2002 and January 2004 were in-luded in the study. Patients with acute coronary syn-

rom the Division of Cardiology, The University Hospital Eppendorf,amburg; Medical Faculty, Pharmacological Haemostaseology, Uni-ersity of Jena, Jena; and Department of Medicine II, Johannes Guten-erg-University Mainz, Mainz, Germany. This study was supported by

he Deutsche Forschungsgemeinschaft (Mu 1079/4-1), Bonn,ermany. Manuscript received June 22, 2004; revised manuscript

eceived and accepted August 24, 2004.Address for reprints: Ascan Warnholtz, MD, Klinikum der Johannes

utenberg-Universität Mainz, 2. Medizinische Klinik Langenbeckstra�e, D-55131 Mainz, Germany. E-mail: warnholtz@2-med.klinik.uni-

rome, decompensated heart failure, ongoing GP IIb/IIa receptor antagonist therapy, or a history ofemorrhagic diathesis were excluded.

Study protocol: Endothelium-dependent, flow-me-iated dilation (FMD) of the brachial artery was non-nvasively examined by 2-dimensional, high-resolu-ion ultrasonic imaging as described previously byevine et al.3 Briefly, 2-dimensional images of the

ight brachial artery and pulsed-Doppler flow velocityignals were obtained with an ATL 7.5- to 12-MHzinear array transducer on an ATL HDI5000 ultra-ound system (Philips, Hamburg, Germany). Imagingas performed in a dark quiet room at a temperaturef 20°C to 23°C. Patients rested in the supine positionor at least 5 minutes before the first scan and re-ained supine until the final recording was acquired.

mages were obtained approximately 5 cm above thentecubital crease. First, baseline 2-dimensional im-ges were acquired followed by pulsed-Doppler bloodow velocity with the signal at a 67° angle to theessel lumen and the 1.0-mm wide gate positioned athe center of the artery. To induce hyperemia, a 3.5-inide blood pressure cuff (D.E. Hokanson, Inc., Bele-ue, Washington) was inflated at the upper arm to 50m Hg above systolic blood pressure or to �200 mmg. Arterial occlusion was kept for 5 minutes with the

ransducer carefully maintained in the identical posi-ion. The cuff was then rapidly deflated and pulsed-oppler velocity signals were recorded for 5 seconds.t 60 seconds after cuff deflation, 2-dimensional im-

ges of the brachial artery were recorded for a period

f 5 seconds.

0002-9149/05/$–see front matterdoi:10.1016/j.amjcard.2004.08.057

dwmewwAaa

awriAtTibutotdtb

pirildsctce

sbb2tt0bib00p

ptufidgpagttS

R

s(dnwtgcp

low-

In a subset of patients, nitroglycerin-induced, en-othelium-independent dilation of the brachial arteryas examined. In these patients, a second baselineeasurement was obtained after 10 minutes of recov-

ry. Subsequently, sublingual nitroglycerin (0.8 mg)as administered, and brachial artery measurementsere obtained after 4 minutes as previously described.ll images were digitally recorded in Digital Imaging

nd Communications in Medicine format for laternalysis.

All patients were examined within 2 hours beforend within 30 minutes after completion of PCI. PCIas performed by standard procedure. In patients who

eceived a coronary stent, a 300-mg clopidogrel load-ng dose was orally given at the end of the procedure.ll patients received an intra-arterial bolus of unfrac-

ionated heparin (5,000 to 10,000 IU before PCI).hirty-three patients received an intravenous bolus

njection of the GP IIb/IIIa receptor antagonist tirofi-an (10 �g/kg body weight), whereas 33 patients whonderwent PCI did not receive tirofiban and served ashe control group. To further evaluate the effect of PCIn endothelial function, in another group of 11 pa-ients with coronary artery disease who underwentiagnostic coronary angiography without coronary in-ervention, FMD of the brachial artery was evaluatedefore and after diagnostic coronary angiography.

Brachial artery image and pulsed-Doppler velocityrofile analysis: Brachial artery diameter was analyzedn a 5- to 15-mm segment before and after induction ofeactive hyperemia. Special care was taken to analyzedentical segments by the identification of anatomicandmarks. The brachial artery segment diameter wasetermined by commercially available edge-detectionoftware (Brachial Analyzer, Medical Imaging Appli-ation, Iowa City, Iowa). The diameter resulted fromhe average of 5 end-diastolic frames. FMD was cal-ulated as the percent change in brachial artery diam-

TABLE 1 Patient Characteristics

VariableTirofiban Group

(n � 33)Control

(n �

Age (yrs) 66 � 2 64 �Women 2 (6%) 3 (Body mass index (kg/m2) 26.8 � 0.6 27.4 �Hypertension 16 (48%) 15 (Smoker 7 (21%) 13 (Diabetes mellitus 5 (15%) 9 (Hypercholesterolemia 21 (64%) 22 (LDL cholesterol (mg/dl) 110 � 6 109 �HDL Cholesterol (mg/dl) 49 � 3 48 �Triglycerides (mg/dl) 140 � 10 151 �Systolic blood pressure (mm Hg) 139 � 3 136 �Diastolic blood pressure (mm Hg) 77 � 1 77 �PROCAM score 11.2 � 1.5 11.8 �Multivessel coronary disease 26 (79%) 25 (ACE inhibitor therapy 16 (48%) 17 (Lipid-lowering therapy 17 (52%) 25 (

Data are expressed as mean � SEM or number (percentage).The PROspective Cardiovascular Munster (PROCAM) study score: calculate

myocardial infarction based on 10-year follow-up in the PROCAM study.4

ACE � angiotensin-converting enzyme; HDL � high-density lipoprotein; LDL �

ter in response to hyperemia. Brachial artery blood a

CORONARY ARTER

flow at rest and during reactive hy-peremia was determined by the aver-age of the flow velocity time integralof the first 3 beats multiplied by ves-sel cross-sectional area and heartrate. The relative increase in bloodflow during reactive hyperemia wasexpressed as the percent increase inflow from baseline. Image analysiswas performed in a blinded manner.

Reproducibility and repeatability:The reproducibility and repeatabilityof the determination of FMD of thebrachial artery have been validatedby numerous previous investiga-tors.3,5,6 The reproducibility of theautomated edge-detection softwarehas also been previously published.7To validate the use of this method inour own laboratory, we examined 20healthy volunteers with no apparentcardiovascular disease. To test thereproducibility of our image analysis

oftware and to assess the intraobserver variability, 1aseline study of each subject was analyzed twice in alinded fashion. Linear regression analysis of the-vessel diameter determinations revealed a correla-ion coefficient of 0.99. The average difference be-ween determinations was 0.034 � 0.008 mm (0.8 �.2% of the vessel diameter). To assess the variabilityetween 2 studies, 10 healthy volunteers were exam-ned twice within 2 hours. The average differencesetween 2 determinations by the same observer were.07 � 0.06 mm for baseline diameter and 0.56 �.57% for FMD. These data agree with previouslyublished methods.8

Statistical analysis: The prospectively determinedrimary end point of the study was the effect ofirofiban on brachial artery FMD. The brachial arteryltrasound parameter was compared between the tiro-ban and the control group using the unpaired Stu-ent’s t test. Comparison of the parameter within theroups before and after PCI was calculated using theaired Student’s t test or Wilcoxon’s signed rank tests appropriate. Clinical characteristics of the 2 studyroups were compared using the unpaired Student’s test, Mann-Whitney rank-sum test, and chi-square or zest as appropriate. All data are expressed as mean �EM.

ESULTSStudy patients: In all, 66 patients were enrolled in the

tudy. Thirty-three patients received a bolus of tirofiban10 �g/kg body weight) for clinical reasons (coronaryissection, thrombus formation, complex bifurcation ste-osis, inability to achieve an optimal balloon inflation),hereas 33 patients who underwent PCI did not receive

irofiban. There were no differences with regard to age,ender, body mass index, blood pressure at rest, bloodholesterol levels, the percentage of current smokers oratients with diabetes, multivessel disease, concomitant

upp Value

0.190.99

7 0.52) 0.99) 0.19) 0.37) 0.99

0.970.980.340.50.7

6 0.93) 0.99) 0.94) 0.1

sk score for acute

density lipoprotein.

Gro33)

29%)

0.45%39%27%67%

7212321.

76%52%76%

d ri

ngiotensin-converting enzyme inhibitors, or statin ther-

Y DISEASE/TIROFIBAN AND ENDOTHELIAL DYSFUNCTION 21

altb

ndtwrbiaatmt

bpetgdowrlv

Ft(ti<

0.4

2

py between the 2 groups. Patient characteristics areisted in Table 1. There were no differences with regardo the interventional conditions of PCI, the time interval

IGURE 1. Effect of GP IIb/IIIa inhibition on brachial artery dila-ion in response to hyperemia. Intravenous injection of tirofiban10 �g/kg body weight) improved FMD of the right brachial ar-ery in patients who underwent PCI, whereas FMD deterioratedn patients who underwent PCI without tirofiban treatment. *p

0.0001 versus pretreatment; †p <0.001 versus tirofiban.

TABLE 2 Characteristics of Percutaneous Coronary Intervention (

VariableTirofiban Group

(n � 33)Control G

(n � 3

Duration of PCI (min) 75 � 4 68 �Direct stenting 8 (24%) 7 (2Stent after balloon angioplasty 22 (66%) 20 (6Balloon angioplasty (no stent) 3 (9%) 6 (1Radiocontrast dye (ml) 217 � 16 199 �Intracoronary heparin (IU) 6,875 � 268 6,812 �Clopidogrel pretreatment 10 (30%) 10 (3Time interval post-PCI to FMD

measurement (min)20 � 3 25 �

Fasting time (h) 16.3 � 0.4 17.0 �

TABLE 3 Brachial Artery Endothelial Function Results

Tirofiban Group(n � 33)

Control G(n �

Baseline brachial arterydiameter (mm)

Pre-PCI 5.06 � 0.08 5.02 �Post-PCI 5.12 � 0.08 5.12 �p Value 0.25 0.1

Reactive hyperemia (% increase)Pre-PCI 785 � 63 681 �Post-PCI 844 � 61 777 �p Value 0.34 0.1

Increase in FMD diameter (mm)Pre-PCI 0.30 � 0.02 0.30 �Post-PCI 0.39 � 0.02 0.23 �p Value �0.0001 0.00

FMD (%)Pre-PCI 6.0 � 0.4 6.1 �Post-PCI 7.8 � 0.5 4.7 �p Value �0.0001 0.00

Data are expressed as mean � SEM

etween the end of PCI and the FMD measurement or t

2 THE AMERICAN JOURNAL OF CARDIOLOGY� VOL. 95 J

frequency of long-term clopidogrelpretreatment, respectively (see Table2).

Brachial artery ultrasound results:Brachial artery parameters for pa-tients in the tirofiban and controlgroups are listed in Table 3. PCI ledto a significant deterioration of FMDof the brachial artery. In contrast,treatment with tirofiban (10 �g/kgbody weight) resulted in a significantimprovement in FMD (Figure 1). Ni-troglycerin-induced, endothelium-in-dependent dilation of the brachial ar-tery remained unchanged and did notdiffer significantly between the 2groups (Figure 2).

Effect of diagnostic coronary an-giography on brachial flow-mediateddilation: In the group of patients whounderwent diagnostic coronary an-giography without PCI, the clinicaland sonographic baseline character-istics were comparable to the groupwho underwent PCI. In contrast toPCI, diagnostic coronary angiogra-phy did not influence FMD (5.4 �0.8% before vs 5.3 � 0.9% aftercoronary angiography, p � 0.94),suggesting a direct inhibitory effectof PCI on peripheral endothelialfunction.

DISCUSSIONThe present study revealed 2 new

and important findings: We demon-strated for the first time that PCI, but

ot diagnostic coronary angiography, causes acuteeterioration of endothelial function in peripheral ar-erial conductance vessels. It is remarkable that thisas demonstrated in the peripheral brachial artery,

emote from the intervention site. Furthermore, tirofi-an was able to inhibit this deterioration and even tomprove endothelial function of the brachial arteryfter PCI. These findings clearly suggest that plateletctivation, per se, is able to impair endothelial func-ion of peripheral conductance vessels, all of whichay explain, at least in part, the beneficial effects of

irofiban in patients undergoing PCI.The deterioration of endothelial function of the

rachial artery after PCI that we observed in our studyoints toward a systemic rather than a local coronaryffect at the intervention site. In addition, the effect ofirofiban on PCI-induced endothelial dysfunction sug-ests that platelet activation is involved in PCI-in-uced endothelial dysfunction. The exact mechanismf how blockade of the GP IIb/IIIa receptor interactsith the endothelium remains unclear. The GP IIb/IIIa

eceptor serves as the common final pathway to plate-et binding to fibrinogen in response to binding ofarious adhesion receptors to constituents of the ex-

)

pp Value

0.1) 0.99) 0.8) 0.48

0.445 0.88) 1.0

0.11

0.17

upp Value

0.750.99

0.20.36

3 0.913 0.0003

0.930.0006

PCI

rou3)

61%0%8%11310%2

ro33)

0.10.11

50392

0.00.08

0.60.76

racellular matrix, such as collagen, von Willebrand

ANUARY 1, 2005

fagsGita

tscsc

rgpgpitmnti

1At2It3An4rc5md6Jd27f28Sp

Fpet

actor, or soluble agents, such as thromboxane A2,denosine diphosphate, and thrombin.9 Because inte-rins are involved in both inside-out and outside-inignaling, improvement in endothelial function due toP IIb/IIIa receptor blockade can be secondary to

nhibition of signaling cascades within the platelet oro the prevention of effects due to platelet adhesionnd aggregation.

Several limitations of our study have to be men-ioned. Patients were not randomly assigned to thetudy groups but treated with tirofiban according tolinical and interventional criteria as decided by thetudy-independent interventionalist. We therefore

IGURE 2. Effect of GP IIb/IIIa inhibition on endothelium-inde-endent dilation in response to 0.8 mg of sublingual nitroglyc-rin. Tirofiban did not change nitroglycerin-induced dilation ofhe right brachial artery.

annot exclude confounding factors due to the lack of9c

CORONARY ARTER

andomization. However, both groups, as well as theroup that underwent diagnostic coronary angiogra-hy, did not differ with respect to clinical and sono-raphic baseline characteristics. Our study does notrovide information with respect to the duration of thempairment of FMD after PCI. However, it is difficulto assess the time course of FMD after PCI, becauseost interventionally treated patients received coro-

ary stents, and implantation of stents requires initia-ion of clopidogrel therapy that may independentlynfluence FMD.

. Aymong ED, Curtis MJ, Youssef M, Graham MM, Shewchuk L, Leschuk W,nderson TJ. Abciximab attenuates coronary microvascular endothelial dysfunc-

ion after coronary stenting. Circulation 2002;105:2981–2985.. Heitzer T, Ollmann I, Koke K, Meinertz T, Munzel T. Platelet glycoproteinIb/IIIa receptor blockade improves vascular nitric oxide bioavailability in pa-ients with coronary artery disease. Circulation 2003;108:536–541.. Levine GN, Frei B, Koulouris SN, Gerhard MD, Keaney JF Jr, Vita JA.scorbic acid reverses endothelial vasomotor dysfunction in patients with coro-ary artery disease. Circulation 1996;93:1107–1113.. Assmann G, Cullen P, Schulte H. Simple scoring scheme for calculating theisk of acute coronary events based on the 10-year follow-up of the prospectiveardiovascular Munster (PROCAM) study. Circulation 2002;105:310–315.. Corretti MC, Plotnick GD, Vogel RA. Correlation of cold pressor and flow-ediated brachial artery diameter responses with the presence of coronary artery

isease. Am J Cardiol 1995;75:783–787.. Sorensen KE, Celermajer DS, Spiegelhalter DJ, Georgakopoulos D, Robinson, Thomas O, Deanfield JE. Non-invasive measurement of human endotheliumependent arterial responses: accuracy and reproducibility. Br Heart J 1995;74:47–253.. Mancini GB, Yeoh E, Abbott D, Chan S. Validation of an automated methodor assessing brachial artery endothelial dysfunction. Can J Cardiol 2002;18:59–262.. Duffy SJ, Keaney JF Jr, Holbrook M, Gokce N, Swerdloff PL, Frei B, Vita JA.hort- and long-term black tea consumption reverses endothelial dysfunction inatients with coronary artery disease. Circulation 2001;104:151–156.

. Leclerc JR. Platelet glycoprotein IIb/IIIa antagonists: lessons learned fromlinical trials and future directions. Crit Care Med 2002;30:S332–340.

Y DISEASE/TIROFIBAN AND ENDOTHELIAL DYSFUNCTION 23