CORONARY ARTERY DISEASE

Contrast Stress Echocardiography for the Diagnosis ofCoronary Artery Disease in Patients With Chest Painbut Without Acute Coronary Syndrome: Incremental

Value of Myocardial Perfusion

Nicola Gaibazzi, MD, Claudio Reverberi, MD, Angelo Squeri, MD, Giuseppe De Iaco, MD, Diego Ardissino, MD,and Tiziano Gherli, MD, Parma, Italy

Background: The inappropriate admission of patients with noncardiac chest pain is an enormous cost to so-ciety. Myocardial perfusion imaging (MPI) could prove effective in the risk stratification of patients in whomacute coronary syndromes are ruled out by electrocardiography and troponin levels, thanks to its incrementalsensitivity beyond that of wall motion (WM) criteria for obstructive coronary artery disease, and still maintainthe excellent safety profile of dipyridamole-atropine stress echocardiography (DASE). The aim of this studywas to test this hypothesis using WM and MPI (WM + MPI) in consecutive patients admitted to a chest painunit.

Methods: Patients presenting to a chest pain unit between January and June 2008 with chest pain and inwhom acute coronary syndromes had been ruled out by normal electrocardiography and cardiac enzymelevels underwent DASE with the addition of contrast MPI. Four hundred consecutive patients were enrolled.

Results: WM + MPI resulted in 71 true-positive findings, compared with 46 by stand-alone WM (P <.05).True-positive results accounted for 46 of 50 positive test results for WM and 71 of 82 positive testresults for WM + MPI (positive predictive value, 92% vs 87%; P = NS). In the subset of patients who un-derwent angiography (n = 116), the sensitivity, specificity, and accuracy for WM compared with WM + MPIwere 63% versus 97% (P < .05), 91% versus 74% (P < .05), and 73% versus 89% (P < .05).

Conclusions: The addition of MPI to standard DASE increased true-positive test results by >50% compared withWM criteria, with a nonsignificant difference in positive predictive value. Twenty-five patients were diagnosedwith obstructive coronary artery disease thanks only to isolated MPI abnormalities; the cardiac origin of theirchest pain would have been mistakenly ‘‘ruled out’’ on the basis of the absence of WM abnormalities. (J AmSoc Echocardiogr 2009;22:404-410.)

Keywords: Stress echocardiography, Contrast, Chest pain, Myocardial perfusion, Dipyridamole

The inappropriate admission of patients with noncardiac chest pain isan avoidable cost to society. The development of safe and accuratediagnostic tools able to rule out obstructive coronary artery disease(CAD) in patients who complain of chest pain of undetermined originwithout electrocardiographic (ECG) or myocardial enzyme changes isa priority. Previous studies showed that predischarge stress echocardi-ography (SE) had independent predictive value on future cardiacevents in patients with chest pain in whom acute coronary syndromeswere ruled out by electrocardiography and troponin measure-ments.1,2

From Dipartimento del Cuore, Azienda Ospedaliero, Universitaria di Parma,

Parma, Italy.

Reprint requests: Nicola Gaibazzi, MD, Dipartimento del Cuore, Azienda Ospeda-

liero, Universitaria di Parma, Via Gramsci 14, 43100 Parma, Italy (E-mail: nicola.

gaibazzi@inwind.it).

0894-7317/$36.00

Copyright 2009 by the American Society of Echocardiography.

doi:10.1016/j.echo.2009.01.017

404

The analysis of myocardial perfusion during dipyridamole-atropineSE (DASE) could prove effective in this subset of patients, thanks to itsincremental sensitivity beyond standard wall motion (WM) criteria forCAD detection, while maintaining the excellent safety profile ofDASE.3,4 The potential increase in sensitivity obtained with contrastmyocardial perfusion imaging (MPI) could be clinically relevant topatients, because MPI analysis has been demonstrated to be capableof predicting mortality and nonfatal myocardial infarctions in patientswith suspected CAD after adjustment for clinical data and WManalysis.5,6

Perfusion defects during vasodilator stress develop because of adecrease in both components of myocardial blood flow, volumeand velocity, in the capillary bed distal to a stenosis.7,8 Although thequantitative measurement of myocardial blood flow reserve in eachmyocardial segment has proved accurate in detecting CAD inselected subsets of patients, it is time consuming and technicallychallenging.

The visual detection of regional discrepancies on MPI during stressis an alternative strategy for CAD detection that may be easier to ap-ply in the clinical setting.9

Journal of the American Society of EchocardiographyVolume 22 Number 4

Gaibazzi et al 405

We prospectively tested whether the addition of real-time MPI afterdipyridamole can increase the sensitivity of DASE for the detection ofCAD in a population of patients admitted to our chest pain unit, clin-ically at low to intermediate risk for CAD, in whom acute coronarysyndromes were ruled out by normal ECG results and troponin levels.The relative values of WM and MPI criteria were compared for accu-racy in a subgroup of patients who underwent coronary angiography.

METHODS

Patient Population

All consecutive patients presenting to our chest pain unit betweenJanuary and June 2008 in whom acute coronary syndromes wereruled out by serial electrocardiography and cardiac enzyme levelsand who met the inclusion criteria were enrolled and underwentDASE within 5 days after their index chest pain episode. The inclusioncriteria were (1) a stress test requested by the cardiology consultantfor an episode of chest pain of suspected cardiac origin, unexplainedby extracardiac conditions; (2) the absence of new ischemic changeson electrocardiography or raised cardiac enzyme levels on $2 serialmeasurements; and (3) low to intermediate pretest risk for CAD, asestimated using a table of risk based on age, gender, chest paintype, and number of risk factors (Diamond and Forrester’s10 criteriaintegrated with Duke database data11). The exclusion criteria were(1) left ventricular ejection fraction (LVEF) < 35%, (2) severe valvulardisease, and (3) frequent or sustained ventricular arrhythmias orhemodynamic instability of any cause.

Study Protocol

All patients with positive results on DASE, on the basis of abnormalWM or MPI results or both, underwent quantitative coronary angiog-raphy (QCA) per protocol; patients with negative results on DASEunderwent QCA only when felt appropriate by the referring physi-cian, on the basis of clinical judgment (Figure 1).

Figure 1 Study design and diagnostic flowchart of patients(pts).

The study complied with the Declaration of Helsinki. Patients gavespecific informed consent when the administration of SonoVue(Bracco Imaging Italia srl, Milan, Italy) occurred < 48 hours afterthe chest pain episode; in fact, this situation is still interpreted as a rel-ative contraindication in the 2008 European Medicines Agency’s12

recommendations for SonoVue administration, even if the presenceof a stable condition at the time of administration seems to be thereal reason for this warning. All patients gave written informed con-sent to the study protocol, which was approved by the institutional re-view board of our hospital.

Clinical Evaluation

All patients were assessed clinically, including history, evaluation ofcardiac risk factor profile, medication use, height, and weight.

Echocardiography

Stress Protocol. Patients underwent standard DASE with adjunc-tive MPI between the end of dipyridamole infusion (0.84 mg/kg over10 minutes) and the beginning of atropine infusion 4 minutes later (at-ropine up to 1.5 mg in 2 minutes). Aminophylline was used to reversethe effect of dipyridamole. Consolidated endpoints and contraindica-tions to DASE were used. Known allergy to sulfonamides, pregnancy,and lactation were considered contraindications to the administrationof echocardiographic contrast media (SonoVue). In cases of contrain-dications to dipyridamole, dobutamine was used as a substitute, andin cases of contraindications to atropine, dobutamine 20 mg/kg/minwas infused after the dipyridamole stage. All patients entered afollow-up program.

Standard and Myocardial Contrast Echocardiography. Pa-tients underwent both WM and MPI studies using an iE33 echocar-diograph with an S5 scan head (Philips Ultrasound, Bothell, WA).MPI was performed activating low–mechanical index (MI) powermodulation imaging after the end of dipyridamole infusion, whileWM acquisition was performed after atropine infusion, at peak heartrate. Flash-replenishment cine loops in the apical 4-chamber, 3-cham-ber, and 2-chamber views were digitally acquired, starting 1 minuteafter the initiation of SonoVue infusion (0.8-1.0 cm3/min) and con-tinuing through the end (4 minutes later, using one vial of SonoVue).A rotating infusion pump was used (BR-INF 100; Bracco Imaging Ita-lia). After the 4 minutes dedicated to MPI, atropine was administeredand SonoVue infusion stopped. Left ventricular opacification gener-ally persisted long enough to allow for peak WM imaging with theleft ventricle still opacified; if not, the residual contrast in the pumptubing (0.8 cm3) was infused at this time by means of a saline bolus.MPI was performed acquiring both triggered and real-time flash-re-plenishment sequences at low MI (0.08-0.12) for each view; triggeredimages were acquired at 1 image per cardiac cycle using ECG gatingat 250 ms after the R wave, while real-time mode acquired images at39 frames/s. High-MI ‘‘flash’’ frames (8 frames; MI = 1.13) were deliv-ered to destroy the microbubbles; on completion of the flashsequence, low-MI imaging automatically resumed. Myocardial con-trast replenishment was visualized, and images were acquired from1 cycle before flash frames through 10 cycles afterward.

Interpretation of WM. Regional WM analysis was evaluated atbaseline and at peak stress by a semiquantitative assessment of theWM score index with the 16-segment model of the left ventricle, ac-cording to the recommendations of the American Society of Echocar-diography.13 Positive test results were defined as the occurrence in $1

406 Gaibazzi et al Journal of the American Society of EchocardiographyApril 2009

segment of either a new dyssynergy in a region with normal rest func-tion or worsening of rest dyssynergy. Resting akinesia becoming dys-kinesia was not considered a criterion for positivity, because this canresult from passive stretching.

Interpretation of MPI. Normal perfusion after dipyridamole wasassigned if myocardium was fully replenished 1.5 to 2 seconds afterthe end of flash impulse; perfusion was defined abnormal if myocar-dium was not replenished after this time but later filled from subepi-cardium to subendocardium.

Basal segments were left out from interpretation, except whenclearly interpretable. Defects that were not at least partly replenishedwithin 10 cycles after flash were discarded as artifacts in patients withno known previous myocardial infarctions.

Defects in patients with previous infarctions, if detected in the cor-responding myocardial segments, were accepted as normal in seg-ments that were akinetic, dyskinetic, or scarred at baseline;otherwise, when segments were not akinetic, dyskinetic, or scarredat baseline, an additional MPI study was performed at recovery to as-sess the reversibility of the defect. The cutoff for normal replenish-ment after aminophylline was considered 4 seconds after thepostflash frame; only if no perfusion defect was found at recoverystage, the results of MPI were finally defined as positive in these pa-tients for inducible defects.

MPI flash-replenishment cine loops were separately evaluated off-line for WM and visual perfusion by two experienced echocardiog-raphers (N.G., C.R.), using commercially available software (Xceleraand QLAB version 6.0; Philips Medical Systems, Best, the Nether-lands). Disagreement was resolved by consensus with a third echocar-diographer (A.S.).

QCA

Only coronary angiograms obtained < 40 days after DASE were con-sidered. QCA was performed by an experienced cardiologist (G.D.),unaware of the results of echocardiography. Any visually evident ste-nosis was measured using a computer-based system for quantitativeanalysis (Qangio XA version 7.0; Medis Medical Imaging, Leiden,the Netherlands) and expressed as percentage narrowing using thenearest normal-appearing region as the reference. CAD was definedas > 50% luminal stenosis in $1 major coronary arteries. Multivesseldisease was defined when >1 of the 3 major coronary arteries (or leftmain trunk) was affected by a > 50% stenosis.

A true-positive result on DASE was defined as a WM or MPI abnor-mality matching the perfusion territory of a > 50% stenosis diagnosedon QCA. Coronary territories were assigned according to the recom-mendations of the American Society of Echocardiography.13

Statistical Analysis

Continuous variables were compared using Student’s t test and arepresented as mean 6 SD. Categorical variables were examinedwith a c2 test when appropriate (expected frequency > 5); other-wise, Fisher’s exact test was used. Interval variables, such as theThrombolysis in Myocardial Infarction risk score, were comparedin unpaired groups using the Mann-Whitney test. Sensitivity, specific-ity, accuracy, and positive predictive value were calculated usingstandard definitions and are presented with 95% confidence inter-vals. Differences between sensitivity, specificity, and accuracy usingWM or WM combined with MPI (WM + MPI) were analyzed usingMcNemar’s test and the Youden index; c2 statistics were used to an-alyze difference in positive predictive value. A P value < .05 (2 sided)

was considered statistically significant. Intraobserver agreement wasdetermined for both WM and MPI in 25 randomly selected patients.Results are presented as percentages with corresponding k values.Statistical analysis was performed using SPSS version 15.0 (SPSS,Inc, Chicago, IL). We had full access to and take full responsibilityfor the integrity of the data. All authors have read and agreed tothe report as written.

RESULTS

The enrollment process and diagnostic flow of patients are summa-rized in Figure 1. Four hundred thirteen consecutive patients metthe enrollment criteria. Thirteen were excluded, 11 because of insuf-ficient echogenicity and 2 because of known allergies to sulfon-amides. Consequently, 400 patients were finally enrolled.

Patient Demographics

Entire Study Group Versus QCA Group. There were no statisti-cally significant differences between the entire study group (n = 400)and the subgroup of patients who underwent QCA (n = 116) regard-ing demographics, number of risk factors, history of previous myocar-dial infarction or percutaneous coronary intervention, and LVEF(Tables 1 and 2).

QCA Subgroup. One hundred sixteen patients (29% of the entirestudy group) underwent quantitative QCA, 82 (71%) because of pos-itive results on DASE (either WM or MPI) and 34 (29%) because ofhigh clinical suspicion for CAD despite negative results on DASE.Among the 73 patients with $1 stenosis > 50%, 22 patients (30%)

Table 1 Baseline clinical, stress echocardiographic, andquantitative coronary angiographic characteristics

Variable All patients QCA group P value

n 400 116 —Age (y) 67 6 10 66 6 10 NS

Men 218 (54%) 64 (55%) NS

$ 2 risk factors* 320 (80%) 94 (81%) NS

Diabetes mellitus 87 (22%) 24 (21%) NSPrior myocardial

infarction or PCI

137 (34%) 41 (35%) NS

Baseline LVEF> 50% 295 (74%) 81 (70%) NSTIMI risk score# 2 321 (80%) 98 (84%) NS

CAD, $ 1 stenosis

> 50%

73 (18%) 73 (63%) <0.05

Single-vessel disease/multivessel disease

— 31 (42%)/42 (58%)

—

Abnormal WM results 50 (13%) 50 (43%) <0.05

Abnormal MPI results 82 (21%) 82 (71%) <0.05

Peak RPP 15,437 6 3,517 14,828 6 3,443 NSD (peak � rest) HR

(beats/min)

38 38 NS

D (peak � rest) BP(mm Hg)

1 �1 NS

BP, Blood pressure; HR, heart rate; PCI, percutaneous coronary inter-

vention; RPP, rate-pressure product; TIMI, Thrombolysis in Myocardial

Infarction.Data are expressed as mean 6 SD or as number (percentage).

*Hypertension, diabetes, current smoking, hypercholesterolemia, family

history of premature CAD, and obesity (body mass index > 29 kg/m2).

Journal of the American Society of EchocardiographyVolume 22 Number 4

Gaibazzi et al 407

had intermediate stenoses (defined as 50%-70%) as the worst steno-sis in their coronary artery trees, while 51 patients (70%) had $1 >70% stenosis (Table 2).

The subgroup with angiographic CAD was significantly older, witha higher prevalence of men and reduced LVEFs compared with thesubgroup without CAD.

Feasibility and Safety

The test could be performed and interpreted for both WM and MPI inall 400 patients finally enrolled, after the initial exclusion of 13 pa-tients for poor acoustic windows or sulfonamide allergies. The feasibil-ity of DASE was 97%.

Among the 400 patients who underwent DASE, there were nomajor complications. In 12 patients (0.03%), significant hypotension(D > 40 mm Hg) developed after dipyridamole infusion, with asso-ciated nausea and true vagal reactions in 6 patients (0.015%), re-versed by aminophylline and the Trendelenburg position; in thesepatients, the tests were in their final phase, and the acquisition ofpeak WM could still be performed while starting aminophylline infu-sion. Mild flushing after dipyridamole was commonly reported(25%). There was no suggestion of any adverse effects related toSonoVue infusion.

Test Results

Fifty dipyridamole-atropine stress echocardiographic studies had pos-itive results for both WM and MPI, and an additional 32 had positiveresults for MPI only. True-positive results accounted for 46 of thetotal of 50 positive test results when only WM was considered (pos-itive predictive value, 92%),and for 71 of the total of 82 positive testresults when WM + MPI was considered (positive predictive power,87%) (P = NS). WM + MPI resulted in 71 true-positive findings,compared with 46 true-positive findings for stand-alone WM (P <.05; Figure 2). Among the additional 25 true-positive findings identi-fied by WM + MPI, 14 patients had single-vessel CAD, and 11 hadmultivessel CAD.

Table 2 Baseline, stress echocardiographic, and quantitativecoronary angiographic characteristics in the subgroup ofpatients who underwent QCA

Variable CAD > 50% No CAD P value

Patients* 73 (63%) 43 (37%) —

Age (y) 68 6 10 64 6 9 <0.05Men 48 (66%) 18 (42%) <0.05

$ 2 risk factors 61 (84%) 34 (79%) NS

Diabetes mellitus 20 (27%) 6 (14%) NS

Prior myocardialinfarction or PCI

27 (37%) 14 (33%) NS

Baseline LVEF > 50% 46 (63%) 35 (81%) <0.05

TIMI risk score # 2 59 (81%) 39 (91%) NS

Abnormal WM 46 (63%) 4 (9%) <0.05Abnormal myocardial

perfusion

71 (97%) 11 (26%) <0.05

Peak RPP 14,548 6 4,320 14,803 6 2,915 <0.05

Abbreviations as in Table 1.

Data are expressed as mean 6 SD or as number (percentage).

*Percentages in this row are calculated referring to the entire QCA group,

while default percentages in all other rows of the table are calculatedinside the specific subgroup (the group with CAD > 50% or the group

with no CAD).

Sensitivity, specificity, and accuracy for WM and WM + MPI, re-spectively, in the QCA group were 63% and 97% (P < 0.05), 91%and 74% (P < .05), and 73% and 89% (P < 0.05) (Table 3, Figure 3).

Intraobserver and Interobserver Variability

The intraobserver agreement for the presence or absence of reversiblevisual perfusion abnormalities or WM abnormalities after full-dosedipyridamole was 96% (k = 0.91) and 100% (k = 1), respectively.The interobserver agreement for the presence or absence of revers-ible visual perfusion abnormalities or WM abnormalities afterfull-dose dipyridamole was 92% (k = 0.82) and 96% (k = 0.90),respectively.

DISCUSSION

Patients in chest pain units selected for stress testing are usually at lowto intermediate pretest risk for CAD, because normal serial ECGresults and troponin levels exclude de facto most intermediate-risk to high-risk subjects, who are directly forwarded to coronaryangiography.

Our study responds to the unmet clinical need for an accurate, safe,radiation-free diagnostic tool to ‘‘rule out’’ obstructive CAD in patientsadmitted to chest pain units without established acute coronarysyndromes.

The main finding of our study is that the addition of stress-only MPIduring standard DASE was not only highly feasible and totally safe butit increased the number of tests with true-positive results for obstruc-tive CAD by > 50%, compared with standard WM criteria. Twenty-five patients were finally diagnosed with $1 coronary stenosis> 50% thanks to isolated MPI abnormalities; the cardiac origin of theirchest pain would have been mistakenly ‘‘ruled out’’ solely on the basisof the absence of WM abnormalities during DASE. Figure 4 shows thetest of a patient finally diagnosed with three-vessel disease who devel-oped isolated MPI abnormalities with normal WM during DASE.Whether the detection of a coronary stenosis by MPI abnormalitieswith apparently normal WM during DASE confers a worse prognosisremains to be investigated in larger cohorts, even if two recent studies,using dobutamine and atropine, suggested this to be the case.5,6

Comparison With Previous MPI Studies

To our knowledge, this study is the first ever to evaluate the incremen-tal effect of MPI beyond WM using low-MI imaging during DASE in

Figure 2 True-positive and false-positive test results in theentire study cohort (n = 400) using WM or WM + MPI analysis.*P < .05 versus WM criteria.

408 Gaibazzi et al Journal of the American Society of EchocardiographyApril 2009

a ‘‘real-life’’ cohort of consecutive patients enrolled from a chest painunit. A few studies addressed MPI utility in chest pain units, butonly one evaluated MPI in conjunction with a stress test14,15; in thatstudy, stand-alone MPI was tested after low-dose dipyridamole(0.56 mg/kg) for prognostic stratification, but the evaluation ofpeak WM abnormalities was not contemplated.14

MPI Incremental Value and Stress Protocol

Early MPI studies were designed to validate MPI compared with sin-gle photon-emission computed tomography, in this case using dipyr-idamole (often 0.56 mg/kg) for stand-alone perfusion analysis.Atropine was not part of the stress protocols, and the presence ofinducible WM abnormalities could not be assessed, because atropineinfusion is crucial for the sensitivity of DASE when using WM as theonly endpoint of the test.16 In recent years, thanks to easier real-timeimaging technology, a few studies have addressed the incrementalsensitivity of MPI beyond WM during dobutamine and atropine,dipyridamole and exercise, exercise testing, or DASE.3,16-19 AlthoughDASE is an accurate and safe protocol for SE, only one study, con-ducted in an extremely selected population (with 86% angiographicprevalence of CAD), addressed the incremental value of MPI duringthis stress protocol.2-4,16,20

Comparison With Standard SE in the Chest Pain Unit

Standard SE may be erroneously perceived as a very sensitive methodto diagnose obstructive CAD. Recent studies in which the sensitivity ofpharmacologic SE was measured reported values ranging from 65% to75% for selected groups of patients undergoing QCA,6,17,20,21 andthese data do not account for the verification bias typical of these stud-ies, which leads to sensitivity overestimation. The main study testingstandard DASE in a chest pain unit demonstrated that it is useful to pre-dict future cardiac events but reported a low prevalence of positive testresults (9%), possibly as a result of the lower risk profile in that studypopulation compared with ours for baseline characteristics.2 The prev-alence of positive test results in our study was indeed only slightlyhigher when considering WM (13%) and increased to 21% onlywhen WM + MPI was considered. We hypothesize that in the afore-mentioned study also, the prevalence of positive test results was lowbecause of the suboptimal sensitivity of WM to detect obstructiveCAD, similar to what happened in our study; the high capability ofWM analysis to predict cardiac events in the cited study is not in con-trast with potential low diagnostic sensitivity for CAD, particularlywhen considering that milder forms of CAD (for stenosis percentagediameter or number of affected vessels) may not necessarily affecthard cardiac events.

Table 3 Diagnostic parameters in the QCA group

Variable WM WM + MPI

Sensitivity

(95% CI)

46/73 (63%) (57%-66%) 71/73 (97%) (92%-99%)*

Specificity

(95% CI)

39/43 (91%) (81%-96%) 32/43 (74%) (66%-78%)*

Accuracy

(95% CI)

85/116 (73%) (66%-77%) 103/116 (89%) (83%-91%)*

CI, Confidence interval.*P < .05 versus WM criteria.

Comparison With Multidetector Computed Tomography(MDCT) in the Chest Pain Unit

Competing imaging techniques, such as MDCT, have shown highersensitivity and negative predictive value for CAD detection in chestpain units compared with standard SE, even if the absence of ischemiaevaluation and the inherent x-ray burden undermine the clinicalusefulness of this otherwise anatomically accurate technique.22

Stress-only MPI, which in our study proved able to increase thesuboptimal sensitivity of WM in the chest pain unit, is consequentlyvery welcome in the clinical arena.

In our study, the sensitivity and specificity of MPI for CAD > 50%,measured on a patient basis, were almost identical to values reportedin one of the few studies using MDCT performed in an intermediate-risk population (and hence comparable with our study).23 In thatstudy, similar to most studies using MDCT, many patients wereexcluded from enrollment, such as those with previous myocardialinfarctions or revascularization, who on the other hand representa consistent proportion of patients presenting to chest pain unitsaround the world. In fact, as many as 34% of patients in our studypopulation had previous infarctions or percutaneous coronary inter-vention.

MPI compares favorably with MDCT for accuracy, withoutx-ray-related safety concerns, without the need for the heavyselection of patients, and with adjunctive functional informationabout the presence of inducible ischemia, which is critical in patientmanagement.

Stenosis Grade

There is no univocal angiographic definition of obstructive CAD, but> 50%, > 70%, and > 75% are the most common definitions. Thestenosis cutoff used to define significant CAD (> 50% in our study)may be relevant to explain the lower than expected sensitivity forstandard WM criteria in our study, potentially higher if only tighter ste-nosis were considered.

Limitations

The use of angiographic data alone ignores the previous occurrenceof coronary events in patients without significant stenosis. Inducible

Figure 3 Diagnostic value for WM with and without MPI fordetection of CAD > 50% in QCA patients (n = 116). *P < .05versus WM criteria.

Journal of the American Society of EchocardiographyVolume 22 Number 4

Gaibazzi et al 409

Figure 4 Assessment of myocardial perfusion after dipyridamole (0.84 mg/kg). From left to right: uniform transmural perfusion seen inapical 4-chamber view before flash, flash image, image taken 3 cardiac cycles (2.1 seconds) after microbubble destruction, andimage taken 6 cycles (4.2 seconds) after destruction. A diffuse subendocardial perfusion defect became clearly apparent after flashimaging, only partially reduced after 6 cycles. Even at peak heart rate, after atropine, this patient had no WM abnormality. Angiog-raphy demonstrated obstructive ‘‘balanced’’ 3-vessel disease.

WM abnormalities or MPI defects in these patients are classified asfalse-positive results. Our protocol excluded patients with severelyimpaired LVEFs (< 35%), since the interpretation of DASE is verysubjective in such patients, in whom other provocative tests may bepreferable. Unbiased clinical comparison between SE and QCA isnot feasible because per protocol performance of QCA in all studypatients would be considered unethical. Our study does not differfrom previous ones because the patients were forwarded to QCAon the basis of either positive results on SE or a perceived high clinicalprobability of CAD despite negative results on SE. In line with previ-ous studies, the prevalence of CAD in our QCA patients was in fact63%; consequently, the results are biased toward the selectionprocess, as usual for this type of diagnostic study.

The accuracy calculated in selected subgroups with high preva-lence of CAD is highly affected by verification bias, overestimatingsensitivity and underestimating specificity.24 We did not report thepositive and negative predictive values of the test derived by sensitiv-ity and specificity obtained in the QCA subgroup, because suchvalues may be interpreted by readers as applicable to the entire group,but this is not the case. The positive predictive value reported in Fig-ure 1 was obtained from data from the entire study group; negativepredictive value cannot be reported, because true-negative andfalse-negative data are not available for the entire group.

QCA subgroup analysis in our study was intended only for thecomparison of WM and MPI relative to diagnostic value, not forabsolute accuracy measurements.

The main unbiased result in our study is the incremental value ofMPI beyond WM for the number of tests with true-positive results.This has direct consequences on sensitivity resulting higher fromMPI, with false-negative results being reciprocally reduced by thehigher number of true-positive results. Nonetheless, sensitivity as anabsolute number for the entire study group could not be calculatedin our study, because the total number of false-negative results re-mained unknown, as usual for this type of clinical study.

CONCLUSIONS

The main finding of our study is that the addition of stress-only MPIduring standard DASE was not only highly feasible and totally safe,but it increased the number of tests with true-positive results forobstructive CAD by > 50%, compared with standard WM criteria.This result was obtained in patients with low to intermediate pretestrisk for CAD, enrolled from the chest pain unit after acute coronarysyndromes were definitely ruled out by serial electrocardiographyand cardiac enzyme measurements. No serious adverse reactionswere noted during or after the tests.

In the selected group of patients who underwent QCA, standardWM demonstrated lower than ideal sensitivity for detection ofCAD (64%), compared with WM + MPI (98%).

Twenty-five patients were finally diagnosed with $1 coronarystenosis > 50% thanks to isolated MPI abnormalities; the cardiacorigin of their chest pain would have been mistakenly ruled out solelyon the basis of the absence of WM abnormalities during DASE.

Whether the detection of a coronary stenosis by isolated MPIabnormalities with apparently normal WM during DASE confersa worse prognosis remains to be investigated in larger cohorts, evenif two recent studies, using dobutamine and atropine, suggested thisto be the case.5,6

REFERENCES

1. Bholasingh R, Cornel JH, Kamp O, van Straalen JP, Sanders GT, Tijssen JG,et al. Prognostic value of predischarge dobutamine stress echocardiogra-phy in chest pain patients with a negative cardiac troponin T. J Am CollCardiol 2003;41:596-602.

2. Bedetti G, Pasanisi EM, Tintori G, Fonseca L, Tresoldi S, Minneci C, et al.Stress echo in chest pain unit: the SPEED trial. Int J Cardiol 2005;102:461-7.

3. Lipiec P, Wejner-Mik P, Krzeminska-Pakula M, Kusmierek J, Plachcinska A,Szuminski R, et al. Accelerated stress real-time myocardial contrast echo-cardiography for the detection of coronary artery disease: comparisonwith 99mTc single photon emission computed tomography. J Am SocEchocardiogr 2008;21:941-7.

4. Varga A, Garcia MA, Picano E; International Stress Echo ComplicationRegistry. Safety of stress echocardiography (from the International StressEcho Complication Registry). Am J Cardiol 2006;98:541-3.

5. Tsutsui JM, Elhendy A, Anderson JR, Xie F, McGrain AC, Porter TR. Prog-nostic value of dobutamine stress myocardial contrast perfusion echocar-diography. Circulation 2005;112:1444-50.

6. Tsutsui JM, Xie F, Cloutier D, Kalvaitis S, Elhendy A, Porter TR. Real-timedobutamine stress myocardial perfusion echocardiography predicts out-come in the elderly. Eur Heart J 2008;29:377-85.

7. Ismail S, Jayaweera AR, Goodman NC, Camarano GP, Skyba DM,Kaul S. Detection of coronary stenoses and quantification of the de-gree and spatial extent of blood flow mismatch during coronary hyper-emia with myocardial contrast echocardiography. Circulation 1995;91:821-30.

8. Wei K, Ragosta M, Thorpe J, Coggins M, Moos S, Kaul S. Non-invasivequantification of coronary blood flow reserve in humans using myocardialcontrast echocardiography. Circulation 2001;103:2560-5.

9. Wei K, Crouse L, Weiss J, Villanueva F, Schiller NB, Naqvi TZ, et al. Com-parison of usefulness of dipyridamole stress myocardial contrast echocardi-ography to technetium-99m sestamibi single-photon emission computedtomography for detection of coronary artery disease (PB127 MulticenterPhase 2 Trial results). Am J Cardiol 2003;91:1293-8.

410 Gaibazzi et al Journal of the American Society of EchocardiographyApril 2009

10. Diamond GA, Forrester JS. Analysis of probability as an aid in the clinicaldiagnosis of coronary-artery disease. N Engl J Med 1979;300:1350-8.

11. Pryor DB, Shaw L, McCants CB, Lee KL, Mark DB, Harrell FE Jr, et al.Value of the history and physical in identifying patients at increased riskfor coronary artery disease. Ann Intern Med 1993;118:81-90.

12. European Medicines Agency. EMEA SonoVue summary of product char-acteristics. Available at: http://www.emea.europa.eu/humandocs/PDFs/EPAR/sonovue/H-303-PI-en.pdf. Accessed June 25, 2008.

13. Pellikka PA, Nagueh SF, Elhendy AA, Kuehl CA, Sawada SG; AmericanSociety of Echocardiography. American Society of Echocardiography rec-ommendations for performance, interpretation, and application of stressechocardiography. J Am Soc Echocardiogr 2007;20:1021-41.

14. Jeetley P, Burden L, Greaves K, Senior R. Prognostic value of myocardialcontrast echocardiography in patients presenting to hospital with acutechest pain and negative troponin. Am J Cardiol 2007;99:1369-73.

15. Tong KL, Kaul S, Wang XQ, Rinkevich D, Kalvaitis S, Belcik T, et al. Myo-cardial contrast echocardiography versus Thrombolysis in Myocardial In-farction score in patients presenting to the emergency department withchest pain and a nondiagnostic electrocardiogram. J Am Coll Cardiol2005;46:920-7.

16. Pingitore A, Picano E, Colosso MQ, Reisenhofer B, Gigli G, Lucarini AR,et al; Echo Persantine (EPIC) and Echo Dobutamine International Coop-erative (EDIC) Study Groups. The atropine factor in pharmacologic stressechocardiography. J Am Coll Cardiol 1996;27:1164-70.

17. Elhendy A, O’Leary EL, Xie F, McGrain AC, Anderson JR, Porter TR. Com-parative accuracy of real-time myocardial contrast perfusion imaging andwall motion analysis during dobutamine stress echocardiography for thediagnosis of coronary artery disease. J Am Coll Cardiol 2004;44:2185-91.

18. Moir S, Haluska BA, Jenkins C, Fathi R, Marwick TH. Incremental benefitof myocardial contrast to combined dipyridamole-exercise stress echocar-diography for the assessment of coronary artery disease. Circulation 2004;110:1108-13.

19. Miszalski-Jamka T, Kuntz-Hehner S, Schmidt H, Hammerstingl C,Tiemann K, Ghanem A, et al. Real time myocardial contrast echocardiog-raphy during supine bicycle stress and continuous infusion of contrastagent. Cutoff values for myocardial contrast replenishment discriminatingabnormal myocardial perfusion. Echocardiography 2007;24:638-48.

20. Picano E, Molinaro S, Pasanisi E. The diagnostic accuracy of pharmacolog-ical stress echocardiography for the assessment of coronary artery disease:a meta-analysis. Cardiovasc Ultrasound 2008;6:30.

21. Tsutsui JM, Elhendy A, Xie F, O’Leary EL, McGrain AC, Porter TR. Safetyof dobutamine stress real-time myocardial contrast echocardiography.J Am Coll Cardiol 2005;45:1235-42.

22. Rubinshtein R, Halon DA, Gaspar T, Jaffe R, Karkabi B, Flugelman MY,et al. Usefulness of 64-slice cardiac computed tomographic angiographyfor diagnosing acute coronary syndromes and predicting clinical outcomein emergency department patients with chest pain of uncertain origin.Circulation 2007;115:1762-8.

23. Hausleiter J, Meyer T, Hadamitzky M, Zankl M, Gerein P, Dorrler K, et al.Non-invasive coronary computed tomographic angiography for patientswith suspected coronary artery disease: the Coronary Angiography byComputed Tomography With the Use of a Submillimeter Resolution(CACTUS) trial. Eur Heart J 2007;28:3034-41.

24. Roger VL, Pellikka PA, Bell MR, Chow CW, Bailey KR, Seward JB. Sex andtest verification bias. Impact on the diagnostic value of exercise echocardi-ography. Circulation 1997;95:405-10.