International Journal of Cardiac Imaging 13: 493–498, 1997. 493c 1997 Kluwer Academic Publishers. Printed in the Netherlands.

Factors influencing the diagnostic accuracy of dobutamine stressechocardiography

K. Schroder1, R. Agrawal2, H. Voller3, B. Kursten2, R. Dissmann2 & H.-P. Schultheiss2

1 REHA-Klinik Ahrenshoop, Dept. of Cardiology, Ahrenshoop, Germany; 2 Klinikum Benjamin Franklin, Dept. ofCardiology, Free University Berlin, Berlin, Germany; 3 Klinik am See, Rudersdorf, Germany

Accepted 14 February 1997

Key words: diagnostic accuracy, dobutamine-atropine, hyperdynamic response, stress echocardiography

Abstract

Background: While Dobutamine stress echocardiography is a well established tool, the range of the diagnosticaccuracy found in the literature is rather large. The main reason for this is the fact, that different test protocolswere used. Aim of this study was to assess the effects of both addition of atropine as well as consideration ofa hyperdynamic response while interpreting the stress echocardiogram on the diagnostic accuracy. Methods andresults: 120 consecutive patients were examined and divided into the following groups: A) achieving their agepredicted heart rate with dobutamine, B) termination of the test due to ischemia, C1) negative test without reachingthe predicted heart rate, and C2) C1 following addition of atropine. All of the echocardiograms were analyzed twice:1) regarding the lack of a hyperdynamic response to dobutamine as ischemia (Hyper analysis), and 2) ignoring thehypercontractility (Conventional analysis). The accuracy of A and B were 88% and 90% resp. Group C1 had a verypoor accuracy of 60%. This rose significantly (p<0.01) after atropine (C2 = 84%), without leading to an increaseof adverse effects. Conventional wallmotion analysis lead to an overall accuracy of 87% (groups A, B, and C2),while Hyper analysis showed an accuracy of 90% (p< 0.01). Conclusions: To achieve a high accuracy Dobutaminestress echocardiography should always be combined with atropine to reach a target heart rate. The wallmotionanalysis should be based on the assumption that a hyperdynamic response to dobutamine is normal, while its lackis indicative of ischemia.

Introduction

Dobutamine stress echocardiography is a wide spreaddiagnostic tool used to detect coronary artery diseasethrough inducing regional wall motion abnormalitiesin the presence of significant coronary artery stenosis.It is most commonly used in patients who are unable toperform an active exercise test, with dobutamine serv-ing as a surrogate for exercise, producing an increasein myocardial oxygen consumption through its positiveinotropic and at high doses also chronotropic effects.By looking at the current literature however, one canfind overall sensitivities ranging from 64% [1] to 96%[2], with specificities from 66% [2] to 95% [3]. Theexplanation for this relatively large range of diagnos-tic accuracy can be found in the protocols used for

administering dobutamine as well as different meth-ods to analyze the wallmotion of the left ventricle.Although 1992 McNeill et al. [4] showed a significantincrease in sensitivity by raising the heart rate to a tar-get level (85% of the age predicted heart rate) throughaddition of atropine at the end of the test, and Pold-ermans et al. [5] proved this new approach to be safe,this amendment to the standard dobutamine protocolis still controversial.

The interpretation and analysis of the wall motionposes another potential source of false results. Dobu-tamine stress testing is often referred to as a substi-tute for active exercise, perfectly imitating the physi-ological hemodynamic response. But while calling ahyperdynamic response of the normal or hypokinet-ic myocardium to exercise a normal, and its lack an

494

ischemic result, this differentiation is not used in dobu-tamine stress echocardiography.

The aim of this study was to assess the effects ofaddition of atropine as well as taking the hyperdynam-ic response into account while interpreting the wallmotion, on sensitivity and specificity for detection ofcoronary artery disease as determined by quantitativecoronary angiography.

Methods

Dobutamine stress echocardiography

After termination of all antianginal medication for atleast 24 hours, dobutamine was administered intra-venously by infusion pump, starting at 10 �g/kg/min.for 3 minutes, and increasing by 10 �g/kg/min. every3 minutes to a maximum of 40 �g/kg/min. As hasbeen recommended by McNeill et al. [4] patients notachieving 85% of the age predicted heart rate (220-age) who had no signs of ischemia, received atropine(starting with 0.25 mg and increasing up to 1.0 mg)intravenously while continuing the maximal dobuta-mine (40 �/kg/min.) infusion.

Standard parasternal and apical views wereacquired at rest using a Toshiba SSH 160 A with 2.5MHz transducer. All echocardiographic images werestored digitally (ImageVue by NovaMicrosonics) atbaseline, following the 10 �g dose, at 40 �g, as wellas after addition of atropine. Wall motion analysis wasperformed visually in a cine-loop fashion using a quad-screen display for side-by-side comparison of the dif-ferent examination stages. The evaluation was basedon the subjective impression of the inward motion ofthe endocardial border towards the center of the leftventricle and the degree of myocardial thickening. Forthis semiquantitative assessment, the left ventricularwalls were divided into 16 segments and the motionscored using a scale ranging from 0 = hyperkinetic to4 = dyskinetic. Test positivity i.e. provocable ischemia,was defined as a new or worsening wallmotion abnor-mality in at least two adjacent segments. All echos wereinterpreted by two experienced investigators (KS, andHV), blinded to the results of the catheterization. Eachecho was analyzed twice, once interpreting the lack ofa hyperdynamic response of the normal or hypokineticmyocardium as ischemia (Hyper analysis), and onceignoring the hypercontractility (Conventional analy-sis).

In our institution inter- and intraobserver variabilityregarding the interpretation of stress echocardiograph-ic examinations is > 90% [6].

Patient characteristics

One hundred twenty consecutive, unselected patientswith known or suspected coronary artery disease wereexamined within 24 hours of a routine cardiac catheter-ization. Indication for examination were chest pain in44 patients with a known, and in 18 patients with asuspected coronary artery disease. The remaining 58patients were examined following an acute myocardialinfarction. Mean age was 57 years (ranging from 39 to76), 86% were male.

Patients were divided into the following subgroups:A) achieving their target heart rate with dobutaminealone, B) showing an ischemic response prior to maxi-mal dobutamine dose (i.e. premature termination of thetest), and C) having a negative result with dobutamineand not reaching the target heart rate. To evaluate theeffect of atropine, group C was evaluated twice, onceat maximum dobutamine dose (hereafter referred to asC1), and once following the addition of atropine (C2).

Coronary angiography

Coronary angiography was performed in all patientswithin 24 hours of the stress test. A lesion> 50% diam-eter stenosis, as determined by quantitative analysis [7]was regarded as significant disease.

Statistical analysis

Results are expressed as mean � SD. Continuos vari-ables were compared using the Student’s t test or 1-way analysis of variance as appropriate. Sensitivityand specificity were calculated using the chi-squareanalysis of contigency tables. Differences between thegroups were compared by using the McNemar’ test. Ap< 0.05 was considered statistically significant.

Results

Study population (Table 1)

Due to poor image quality the dobutamine stress testcould not be performed in 17 patients. Of the remain-ing 103 patients, 4 had adverse events which lead to

495

Table 1. Patient characteristics of the groups A (target Heart rate without atropine), B (prematurediscontinuation), and C (end of test at max. dobutamine dose prior to reaching target heart rate).

A B C

n 43 19 37

Mean Age 59.6 57.6 55.9

Male (%) 33 (77) 16 (84) 30 (82)

suspected CAD (%) 6 (14) 1 (5.3)� 4 (10.5)

known CAD (%) 14 (32.5) 11 (57.9) 11 (29)

AMI /%) 23 (53.5) 7 (36.8) 22 (60.5)

no CAD (%) 6 (14) 1 (5.3)� 4 (11)

SVD (%) 22 (51) 8 (41) 17 (46)

MVD (%) 15 (35) 10 (53)�� 16 (43)

rest HR (bpm) 75.7� 12.9 74.9 � 12.9 69.2 � 9.9

rest systolic BP (mm Hg) 126.2 � 15.3 118.7 � 15.1 118.8 � 10.9

rest Double Product 9986� 3124 7078 � 4162 8247 � 1605

CAD= coronary artery disease, AMI= acute myocardial infarction, SVD= single vessel disease,MVD= multi vessel disease, HR= heart rate, BP= blood pressure. � = p< 0.05 as compared to A,�� = p< 0.05 as compared to A and C.

premature termination of the test (2 non sustained ven-tricular tachycardia, 1 hypotension, and 1 episode ofongoing bigeminus). The study population with evalu-able stress echocardiograms therefore consisted of 99patients. 43 patients reached the target heart rate withdobutamine infusion alone (group A), while 37 did not(group C). Premature termination due to test positivi-ty was observed in 19 patients (group B). Quantitativecoronary angiography revealed a significant stenosis in88 patients (89%). The baseline characteristics of thethree groups are presented in Table 1. The groups didnot differ regarding the gender and age distribution,as well as the hemodynamic baseline data. There wasalso no difference regarding the incidence of singlevessel disease, while group B had more multi vesseldisease than the other two. The indication to performthe test was the same in all groups regarding knowndisease and acute myocardial infarction, while therewere more patients with a suspected disease in groupA.

Hemodynamic changes (Table 2)

There was no significant difference in the baselineheart rate, blood pressure, and consequently also dou-ble product in the four groups (A to C2). The stressinduced changes of these parameters however variedsignificantly. In group A both the mean resting heartrate as well as the systolic blood pressure increased by83% and 24% resp. (p< 0.0001 and p < 0.05 resp.)during the dobutamine infusion (76 � 13 bpm to 139

Table 2. Stress induced changes of the hemodynamic parameters.

HR (bpm) systolic BP Double Product

(mm Hg)

A Rest 76 � 13 126 � 15 9986 � 3124

Max 139 � 13� 156 � 21��� 21734 � 3901�

Difference 83% 24% 118%

B Rest 75 � 13 119 � 15 7078 � 4162

Max 106 � 19�� 147 � 16��� 12072 � 6951��

Difference 41% 24% 71%

C1 Rest 69 � 10 119 � 11 8247 � 1605

Max 110 �17�� 148 � 23��� 16480 � 4286��

Difference 59% 24% 99%

C2 Rest 69 � 10 119 � 11 8247 � 1605

Max 140 � 13� 153 � 21��� 21358 � 3713�

Difference 103% 29% 159%

HR = Heart rate, BP = blood pressure, Rest = baseline, Max = peakpharmacological stimulation. � = p< 0.0001 as compared to baseline,�� = p< 0.01 as compared to baseline, ��� = p< 0.05 as comparedto baseline.

� 13 bpm, and 126 � 15 mm Hg to 156 � 21 mmHg) The changes regarding the double product werealso significant (9986� 3124 compared with 21734�3901, p< 0.0001). Similar reactions to the pharmaco-logical stress were seen in group C2 after addition ofatropine. Heart rate increased by 88% from 69 � 10bpm to 140 � 13 bpm, blood pressure rose 31% from119 �11 mm Hg to 153� 21 mm Hg, and the doubleproduct changed by 148% from 8247�1605 to 21358� 3713 (p< 0.0001, p <0.05, and p <0.0001 resp.)

496

Table 3. Overall sensitivity, specificity, and diag-nostic accuracy.

A B C1 C2

Sensitivity 89 94 58 85

Specificity 86 n.a. 75 75

Diagnostic accuracy 88 90 60 84

n.a.= not available, since a coronary artery diseasecould only be excluded in one patient.

Figure 1. Sensitivity to detect a significant coronary artery stenosisin the different groups. SVD = single vessel disease, MVD = multivessel disease.

The changes registered in groups B and C1 were alsosignificant, but overall not as pronounced as in groupA or C2.

Sensitivity and specificity (Table 3, Figure 1 and 2)

Of the 37 patients with a significant coronary stenosisin group A, 32 had an ischemic response to dobuta-mine (sensitivity of 89%). 19 of the 22 patients witha single vessel disease and 14 of the 15 with a multivessel disease were correctly identified (sensitivity of86% and 93% resp.). The specificity in this group was86%. Group C1 had a poor sensitivity for the entiregroup (58%, 19/37) as well as for the subgroups singleand multi vessel disease (47%; 8/17, and 67%; 11/16resp.). After addition of atropine the sensitivity wassignificantly increased (p < 0.01) to the level of theother groups (85%; 76%; 94% resp.). The specificitywas not influenced by the atropine addition: 75% inboth C1 and C2.

Group B had the best overall sensitivity of 94%(17/19 patients). This was also true for single and multivessel disease (88%; 7/8, and 100%; 10/10). Specifici-ty of this group was not determined, since only onepatient did not have a coronary artery disease.

Figure 2. Sensitivity, specificity, and diagnostic accuracy of theentire patient population divided according to the analysis methodused. Conv = ignoring the hyperdynamic response to dobutamine,Hyper = regarding a hyperdynamic response as normal, its lack aspathological response.

Table 4. Adverse effects in the two groups reaching thetarget heart rate.

A C2 p

Discomfort 8 7 ns

Palpitation 4 3 ns

Angina 3 4 ns

ST changes 2 1 ns

Dyspnoea 1 1 ns

SVT 1 0 ns

PVC 1 3 ns

Total 20/43 19/37

ST = ST-Segment, SVT = supra ventricular tachycardia,PVC = premature ventricular contractions.

After showing the beneficial effect of atropine addi-tion, all of the echocardiograms of groups A, B, andC2 were reanalyzed with respect to hypercontractility(Figure 2). The primary analysis, leading to the resultsstated above, was performed as Conventional analysis,i.e. ignoring the presence or absence of a hyperdynamicresponse to dobutamine/atropinestimulation. This leadto an overall sensitivity of 89% and an overall speci-ficity of 85%. Interpreting the lack of a hyperdynamicresponse as ischemia (Hyper analysis), comparable tonew or worsening dyssynergies, lead to a significantincrease of sensitivity up to 95% (p< 0.01) with onlya very minor, non significant decrease of specificity(83%).

497

Adverse effects (Table 4)

47% of the patients (20/43) in group A had mild tomoderate adverse effects during the test, while 51%(19/37) in group C2 reported the stated events. Therewere no serious events leading to premature termi-nation of the tests and/or requiring specific therapy.All of the events ceased spontaneously. Overall therewas no difference in adverse effect, regardless of thestimulation necessary (dobutamine alone vs. dobuta-mine/atropine) to reach the age predicted target heartrate.

Discussion

Dobutamine stress echocardiography is a relativelynew, but non the less, well established noninvasivediagnostic tool used to detect and evaluate coronaryartery disease. Several studies have shown a very goodcorrelation to angiographic [2, 3, 6] and scintigraph-ic [8] findings resulting in an overall good diagnosticaccuracy. The range of the reported sensitivities andspecificities however is relatively large (64% [1] – 96%[2] and 66% [2] – 95% [3] resp.). The main reason forthis being the fact, that different test protocols wereused during the studies. Two factors seem to be mostinfluential concerning the correlation to the invasiveand nuclear studies: 1) reaching an age predicted tar-get heart rate at the end of the test, and 2) interpretationof wallmotion changes induced by dobutamine stressechocardiography.

Although 1992 McNeil et al. [4] showed that theaddition of atropine to dobutamine in patients who donot reach 85% age predicted heart rate and whose testresults are negative (i.e. not ischemic) during dobuta-mine alone, safely increases the diagnostic accuracyof the test, the use of additional atropine at the end ofa negative dobutamine stress test is still not routinelyused in many centers.

Our study confirmed the data of McNeil et al. [4]proving the beneficial effect of atropine addition at theend of a negative dobutamine stress echocardiogramonthe diagnostic accuracy as determined by comparisonto quantitative coronary angiography. Patients reach-ing their target heart rate with dobutamine alone had anoverall diagnostic accuracy of 88%, which was compa-rable to that of the group of patients with premature ter-mination of the test (B) due to ischemia (90%), as wellas to data reported in the literature. Those patients whodid not reach their target heart and were not ischemic

following the test on the other hand, had a diagnosticaccuracy which was significantly lower than the othergroups (60%). After addition of atropine this value sig-nificantly increased up to the level of the other groups(85%).

The hemodynamic data of the groups was not dif-ferent at baseline. The main difference between thegroups seen throughout our study was the effect of thepharmacological stimulation on the heart rate. Whilein group A the heart rate increased by 83% followingdobutamine stimulation, group C1 just showed a riseof 51%. After atropine addition the heart rate increasedby the approximately same amount (88%) as in groupA. In the ischemic group (B) the positive chronotropiceffect (increase of 41%) was also not as pronounced asin A or C2. This was due to the premature terminationof the test before an adequate rise of the heart rate couldoccur. As was to be expected, the change concerningthe systolic blood pressure was the same for all of thegroups.

As Poldermans et al [5] showed 1994, this proto-col amendment proved to be safe and did not lead toan increase in mild to moderate adverse effects (47%group A vs. 51% group C1) in 99 consecutive, unse-lected patients scheduled for routine stress testing pri-or to a diagnostic catheterization. Non of the patientsincluded in this study had serious adverse events.

Dobutamine is often referred to as an exercise sim-ulation agent, used for patients who are unable to per-form an active exercise (treadmill or bicycle) test. Butwhile the interpretation of exercise stress echocardio-graphy is based on the assumption that the normalresponse to exercise is an increased systolic thickeningand inward motion of the ventricular walls, whereas areduction or no change is indicative of significant coro-nary artery stenosis, this approach is not a commonanalysis concept with dobutamine stress echocardiog-raphy.

We could show, that basing the wallmotion analy-sis of dobutamine stress echocardiographyon the sameassumptions as exercise echocardiography, (i.e. a nor-mal response to graded dobutamine infusion is a hyper-dynamic reaction of the ventricular walls, and lackof hyperdynamic wallmotion is comparable to a newdyssynergies) leads to a significant (p< 0.01) improve-ment of the sensitivity (89% vs. 95%) with just a minor,non-significant reduction in specificity (85% vs. 83%).The diagnostic accuracy of the test increased signifi-cantly from 87% to 90% (p<0.01) using this analysisconcept.

498

We feel that precise definitions of normal andabnormal responses to pharmacological stimulation aswell as uniform protocols of drug application are anecessity to increase the accuracy of the test as wellas to facilitate training and reproducibility. Thereforewe would recommend a routine addition of atropine toevery patient not reaching the age predicted target heartrate at the end of a negative dobutamine stress test, aswell as always taking the hyperdynamic response, orits lack, to dobutamine stimulation into account whileinterpreting the wallmotion at maximum pharmacolog-ical stress.

Acknowledgments

The authors wish to thank Mrs. E. Tessun for the workin the stress echocardiography laboratory as well as theteam of the cardiac catheterization laboratory for theexcellent collaboration.

References

1. Mazeika PK, Nadazin A, Oakley CM. Dobutamine stressechocardiography for detection and assessment of coronaryartery disease. J Am Coll Cardiol 1992; 19: 1203–11.

2. Marcowitz PA, Armstrong WF. Accuracy of dobutamine stressechocardiography in detecting coronary artery disease. Am JCardiol 1992; 69: 1269–73.

3. Cohen JL, Greene TO, Ottenweller J, Binenbaum SZ, WilchfortSD, Kim CS. Dobutamine digital echocardiography for detect-ing coronary artery disease. Am J Cardiol 1991; 67: 1311–8.

4. McNeill AJ, Fioretti PM, El-Said M, Salustri A, Forster T, Roe-landt J. Enhanced sensitivity for detection of coronary arterydisease by addition of atropine to dobutamine stress echocar-diography. Am J Cardiol 1992; 70: 41–46.

5. Poldermans D, Fioretti P, Boersma E, Forster T, Urk H, CornelJ, et al. Safety of dobutamine-atropine stress echocardiographyin patients with suspected coronary artery disease. Am J Cardiol1994; 73: 456–459.

6. Schroder K, Voller H, Dingerkus H, Munzberg H, Dissmann R,Linderer T, et al. Compardison of the diagnostic potential of fourechocardiographic stress tests shortly after acute myocardialinfarction: submaximal exercise, transesophageal atrial pacing,dipyridamole, and dobutamine. Am J Cardiol 1996; 77: 909–914.

7. Linderer T, Wunderlich W, Backs B, Noering J, Schroder R.Edge detection in quantitative coronary arteriography (QCA):The impact of image zoom on edge positioning, accuracy, pre-cision and discrimination. Comp Cardiol 1992; 99–102.

8. Marwick T, Willmart B, D‘Hondt AM, Baudhuin T, Wijns W,Detry JM, et al. Selection of the optimal nonexercise stressfor the evaluation of ischemic regional myocardial dysfunctionand malperfusion: comparison of dobutamine and adenosineechocardiography and 99m Tc-MIBI single photon emissioncomputed tomography. Circulation 1993; 87: 345–354.

Address for correspondence: Klaus Schroder, REHA-Klinik Ahren-shoop, Dept. of Cardiology, Dorfstr. 55, 18348 Ostseebad Ahren-shoop, Germany.Tel: (038220) 63333; Fax: (038220) 63243