prevalence and functional significance of transient st-segment depression during daily life...

Prevalence and functional significance of transient ST-segment depression during daily life activity: Comparisons of ambulatory ECG with stress redistribution thallium 201 single-photon emission computed tomographic imaging

To assess the prevalence and functional significance of ischemic ambulatory ECG responses, we prospectively performed ambulatory ECG monitoring in 244 patients (mean age 61 + IO years) referred for stress redistribution thallium 201 myocardial perfusion scintigraphy. The prevalence of ST-segment depression during ambulatory ECG was 33% among patients with a positive exercise ECG, but prevalence varied in selected patient subgroups. Among three groups with coronary artery disease (CAD), the group with ambulatory ECG ischemia (group 1) had a greater frequency of ischemic thallium responses (p = 0.07), a greater median number of reversible thallium defects (p < 0.05), and a greater summed thallium “reversibility” score (p < 0.05) than did the group with a positive exercise ECG but negative ambulatory ECG response (group 2) or that with negative exercise and ambulatory ECG responses (group 3). Exercise ST depression in group 1 versus group 2 was significantly greater (p = 0.002), occurred at a lower heart rate threshold (p = 0.002), and lasted longer after exercise (p = 0.001). Notably, one third of group 1 patients also manifested evidence of transient ischemic dilation of the left ventricle after exercise (p < 0.01 vs groups 2 and 3), a sign of severe ischemia. However, although functionally less “sick” than group 1 patients, 66% of group 2 patients and 50% of group 3 patients still had an ischemic thallium response, which was sometimes severe. Thus transient ischemia during ambulatory ECG monitoring identifies a functionally sicker cohort of patients with CAD and occurs in approximately one third of CAD patients with positive results of exercise tests. A negative ambulatory ECG response, however, does not exclude functionally significant disease among CAD patients. These results imply that caution should be applied in the interpretation of a negative ambulatory ECG response for the purpose of patient risk stratification. (AM HEART J 1993;125:1247.)

Jacob Klein, MD, Erwin A. Rodrigues, MD, Daniel S. Berman, MD, Florence Prigent, MD, Susan Y. Chao, MS, Tamara Maryon, PhD, and

Alan Rozanski, MD New York, N.Y., and Los Angeles, Calif.

From the Division of Cardiology and Department of Nuclear Medicine, Ce- dars-Sinai Medical Center and the Department of Medicine, University of Los Angeles, School of Medicine; and the Division of Cardiology, Depart- ment of Medicine, St. Luke’+Roosevelt Hospital Center and the Depart-

ment of Medicine, Columbia University College of Physicians and Surgeons.

Funded by the John D. and Catherine T. MacArthur Foundation and the KROC Foundation. Performed while Dr. Klein served as Save a Heart Foundation fellow in preventive cardiology.

Received for publication Sept. 21, 1992; accepted Nov. 6, 1992.

Reprint requests: Alan Rozanski, MD, Division of Cardiology, St. Luke’+ Roosevelt Hospital Center, 114th Street and Amsterdam Avenue, New York, NY 10025.

Copyright ’ 1993 by Mosby-Year Book, Inc.

0002.8703/93/$1.00+.10 4/l/44921

Because of the prognostic relationship between myocardial ischemia and cardiac events1 the assessment of myocardial ischemia has become a central compo- nent in the clinical treatment of patients with coronary artery disease (CAD). Ischemia is most com- monly and generally first measured by its laboratory induction during exercise ECG. Radionuclide stress testing2-4 and ambulatory ECG5-8 both add incre- mental information to exercise ECG for patient risk stratification. For this reason and because of differences in cost, approach, and ease of performing these two tests, comparisons between them are warranted.

1247

1248 Klein et al. May 1993

American Heart Journal

So far, however, the systematic comparison of ambulatory ECG measurements of ischemia with those measured by radionuclide imaging have been relatively limited. Previously we used Bayesian theory to predict the potential strengths and weaknesses of ambulatory ECG as a clinical monitoring technique in patients with suspected or known CAD.g-ll For instance, on the basis of Bayesian theory, there should be a limitation to interpreting positive ambulatory ECG responses when the pretest probability of CAD is low (frequent false-positive responses) and to

interpreting negative ambulatory ECG responses when the pretest probability is high (frequent false- negative responses). ‘J l1 Because ambulatory ECG monitoring has generally been promulgated as a test for prognosis, patients with a high likelihood of or with known CAD are the group to whom the test is usually applied. Thus understanding the potential limitation(s) of ambulatory ECG in this subgroup would be clinically important.

Thus we compared the results of ambulatory ECG monitoring and stress redistribution thallium 201 myocardial perfusion scintigraphy among patients having a high pretest probability of CAD. Using thallium scintigraphy as an independent marker of myocardial ischemia, we evaluated the relationship between positive and negative ambulatory ECG responses and the presence and magnitude of inducible myocardial hypoperfusion.

METHODS Patient population. We prospectively recruited 244 pa-

tients, referred to our laboratory for stress redistribution single-photon emission computed tomographic imaging (SPECT) thallium 201 myocardial scintigraphy, for 24 to 48 hours of ambulatory ECG monitoring. Thallium testing was performed for either diagnostic evaluation (93 patients) or functional assessment of known CAD (151 patients). The latter group included 74 patients with prior myocardial infarction, 45 patients with prior angioplasty, and 32 patients with prior coronary bypass surgery. There were 198 men and 46 women, with a mean age of 63 t 10 years (range 35 to 85 years). Patients unable to exercise and those with potentially uninterpretable ambulatory ECGs were excluded because of significant resting ST segment depression, conduction abnormalities, left ventricular hypertrophy, or digoxin therapy. For relevant analyses, patients were subgrouped on the basis of CAD probability, determined according to the method of Diamond and For- rester,r2 with Bayesian analysis of patient age, sex, symp- toms, risk factors, and results of exercise ECG.

Exercise testing. Maximal exercise treadmill testing was performed by the standard Bruce protocol. The ECG was monitored continuously in three leads (aVF, VI, Vs), and 12-lead ECG was recorded at rest, during each minute after exercise, for at least 5 minutes or until all exercise-in-

duced ECG changes disappeared. The exercise ECG ce- sponse was considered positive if horizontal or downslop ing ST segment depression of at least I mm or of at least 1.5 mm upsloping depression occurred in any lead compared with the baseline value, measured at 0.08 seconds after the J-point. Blood pressure was recorded with a cuff sphygmomanometer at rest, during each stage of exercise, at peak exercise, and at 2-minute intervals after exercise. Our endpoints for premature termination of exercise by the supervising physician are severe angina, high-grade ventricular arrhythmia, severe hypertension, or exertional hy- potension. All patients referred for stress redistribution thallium scintigraphy were routinely instructed to with- hold P-blocking medications for at least 48 hours, calcium channel blocking agents for 24 hours, and long-acting nitrates for 6 hours before testing; 231 patients (95”; ) were tested without anti-ischemic medications, and only 13 patients (5’; J were tested while receiving anti-ischemic medications. This proportion is representative of the com- pliance rate for acute drug cessation before stress testing in our laboratory.

SPECT thallium 201 scintigraphy. At near maximal exercise, 3 to 4 mCi of thallium 201 was injected intravenously and exercise was continued for an additional minute. Im- aging was begun approximately 7 minutes after the injec- tion of thallium with the acquisition of a 5-minute planar anterior view. Tomographic images were then obtained by rotating of the camera. Thirty-two projections were obtained along a circular 180-degree arc, extending from the 45-degree right anterior oblique projection to the left posterior oblique projection. The stress and 4.hour studies were obtained with 40 seconds per projection. If a stress defect was persistent at 4 hours, late (24 to 48 hours) redistribution tomographic imaging was performed, with 60 seconds for image acquisition in each projection, Repeated planar images were also obtained at 4 hours. A large field- of-view scintillation camera containing 75 photomultiplier tubes and a 0.25-inch (0.64 cm) sodium iodine crystal was used, equipped with a low-energy, all-purpose parallel hole collimator. A 20’; energy window centered on the 80 keV peak and a lo’*, window centered on the 167 keV peak were employed. All projection images were stored on magnetic disk with a 64 X 64. 16-bit matrix. Each of the 32 projection images was corrected for nonuniformity with a 30 mil- lion count image of a uniform cobalt 57 flow source. The mechanical center of rotation was determined from the projection data to align the detector data with respect to the reconstruction matrix. Filtered back projection was performed on each clinical data set with a low-resolution Butterworth filter with a cutoff frequency of 0.45 cycles of pixels, order 5, to reconstruct transverse axial tomograms, each 6.2 mm thick, encompassing the entire heart. Vertical and horizontal long-axis tomograms and short-axis slices were extracted from the filtered transaxial tomograms by performing a coordinate transformation with the appro- priate interpolations. No attenuation or scatter corrections were used.

For visual interpretation by the physician, all short- and long-axis tomograms were displayed on transparency film,

Volume 125, Number 5, Part 1

American Heart Journal Klein et al. 1249

Apical

SHORT AXIS

Mid-Ventricular Basal

VERTICAL LONG AX/S

1 2 6

69 3 5 4

7

0 12

0 9 11

10 16

19

65 20

Fig. 1. Diagrammatic presentation of thallium 201 SPECT images. Short-axis slices are divided into three regions, as indicated, and vertical long axis is used to evaluate apex. Twenty myocardial segments are analyzed per patient, as shown.

with the intensity of each image normalized to the maximal pixel value in that image. Segmental scores were assigned to six evenly spaced regions in the apical, midventricular, and basal cuts of the short-axis views and the two apical (inferoapical and anteroapical) segments of the midverti- cal long-axis cut (Fig. 1). A 4-point scoring system was used to assess each segment as follows: 0 = no decrease in regional thallium 201 uptake, 1 = mild decrease, 2 = moderate decrease, and 3 = severe decrease.13 An initial poststress perfusion defect was considered present when a myocardial segment had an initial poststress score of at least 2. Stress defects with a score of 1 or less at 4 or 24 hours redistribution imaging were deemed reversible; stress defects with 4- and 24-hour scores of 2 or greater were deemed nonreversible. A thallium test result was considered positive if two or more of the 20 myocardial segments showed evidence of reversibility. If only nonreversible defects were present, the thallium study was considered negative for the presence of ischemia. To assess the global magnitude of inducible ischemia, we calculated a summed thallium reversibility score, defined as the sum of the scores in the 20 segments after stress minus the sum of the scores at 4 hours, or at 24 hours if further redistribution occurred. The im- mediate poststress and redistribution anterior view planar thallium images were used to identify two additional scin- tigraphic abnormalities: (1) transient ischemic dilation of the left ventricle (TID) after stress and (2) abnormal pulmonary thallium uptake. TID was considered to be present after stress when there was a definite reduction in the size of the left ventricle between the stress and the 4-hour redistribution images.14 The amount of lung uptake of thallium was graded according to a 4-point system: 0 = no increase, 1 = mild or moderate increase, 2 = moderate increase, and 3 = marked increase (approaching intensity of myocardial thallium activity), with disappearance at the time of redistribution imaging.15 Scores of 2 or more were considered abnormal.

Ambulatory ECG. All patients underwent 24 hours, and 66 patients underwent 48 hours, of continuous ECG monitoring within a median of 0 days after thallium exercise

testing; 90% of patients were tested within 7 days of the thallium test. As with thallium testing, anti-ischemic medications were also withheld by patients before and during the period of ambulatory ECG, according to the same reg- imen described for thallium testing. A Cardiodata AM re- corder (MK4, frequency response of 0.05 to 100 Hz) was used, with automatic calibration of the EGG signal with a 1 MV = 10 mm signal at the start of the recording. After careful skin preparation two pairs of bipolar pregelled electrodes were attached and taped to the precordium. The exploring electrodes were attached to Vs, and a modified inferior position and the two indifferent electrodes were attached to the right side of the manubrium. A ground electrode was placed in all patients on the right lower ribs.16 Patients were instructed to carry out their usual daily activities and to fill out a structured diary.i7

The calibrated 24-hour ambulatory monitor tapes were visually analyzed by an experienced technician at 60 times real time with a Cardiodata MK 4 computer. An ambulatory response was considered positive if significant ST segment depression occurred, defined as horizontal or downsloping ST segment depression of more than 1 mm, (or >1.5 mm, if upsloping), less than the isoelectric baseline, measured at 0.08 seconds after the J-point, and persistent for at least 60 seconds. Separation of one episode from the next required that the ECG return to baseline level for at least 3 minutes after an earlier episode. Base- line and representative strips for each ischemic episode were saved and printed out in real time to be read blindly by two independent physicians. Disagreements were set- tled by consensus interpretation of the two observers.

To test the accuracy of our ambulatory ECG system for measuring ST-segment depression compared with 12-lead ECG, 25 randomly selected patients underwent simulta- neous maximal treadmill exercise testing (with a Marque- tte Case 12 system) and ambulatory ECG monitoring. These patients had a high CAD probability and were referred for exercise thallium 201 testing for further exam- ination; the aforementioned exclusion criteria were applied. An exploring electrode positioned at the Vs-like lead

1250 Klein et al.

A. 8.

May 1993


: 4.0

-5 3.0

z 2.0

e ,o

I .o

a 3 0

% -1.0

:, -2.0

5 -3.0

‘74 g -4.0

,h -5.0

; -6.0 cn

-6.0-5.0-4.0-3.0-2.0-1.0 0 1.0 2.0 3.0 4.0

ST Depression on Exercise ECG (mm) ST Depression on Exercise ECG hm)

Fig. 2. A, Comparisons of magnitude of ST segment at peak exercise. B, Magnitude of change in ST segment from resting to peak. Values are in millimeters, measured for ambulatory ECG (each vertical axis) and exercise ECG (each horizontal axis) during exercise treadmill testing in 25 randomly selected patients. Each point represents at least one paired observation. Diagonal line represents line of identity. Close agreement between ambulatory ECG results and simultaneously recorded exercise ECG results was found.

was split through a common junction box, so that the same bipolar signal was read by both the 12-lead ECG and the ambulatory ECG system. Real time ECG strips were taken simultaneously on both ECG systems with the event marker, at rest and at peak exercise. All strips were read in a blinded manner.

Fig. 2 shows the comparison of the magnitude of the ST segment at peak exercise and the magnitude of the change in the ST segment from rest to peak exercise as measured simultaneously by ambulatory and exercise ECG. The ambulatory ECG correlation coefficient (rc) was used as an index of agreement.18 The ambulatory ECG provided ex- cellent agreement with the magnitude of the ST segment at peak exercise and the magnitude of the change in ST segment recorded by the exercise ECG (rc = 0.99, 95% confidence interval = [0.97, 0.991; rc = 0.99, 95% confidence interval = [0.98, 1.001, respectively).

Statistical analysis. Data were expressed as mean _t SD or as proportions. Unevenly distributed data were expressed as median and range. Univariate analysis of cate- goric variables was performed with chi-square analysis or Fisher’s Exact Test. When two groups were compared, quantitative variables were analyzed with the Student t test or Wilcoxon rank-sum test. When more than two groups were compared, quantitative variables were analyzed with analysis of variance of the Kruskal-Wallis test. All significance testing was done at the 0.05 level.

RESULTS

The prevalence of ST segment depression during ambulatory ECG for the total patient population and various clinical subgroups is listed in Table I. Among the 55 patients having a positive ambulatory ECG during 24-hour monitoring the mean duration of ST- segment depression was 19 + 18 minutes and the

median number of episodes was two (range one to seven). Among the 82 patients with a negative exercise ECG, the prevalence of ST-segment depression during 24-hour ambulatory monitoring was rare (2%). By contrast, in 53 of the 162 patients with a positive exercise ECG (33%) transient ST segment depression developed during 24-hour ambulatory monitoring. When patients with a positive exercise ECG response were further subdivided on the basis of general clinical characteristics, the percentage of patients with a positive ambulatory ECG response varied considerably but never reached 50 % in any of the clinical subgroups in Table I, including the patients with typical angina and a positive exercise ECG response. However, when patients were subgrouped on the basis of the magnitude of exercise ECG responses (Fig. 3), cohorts with a high prevalence of positive ambulatory ECG responses could now be identified. The prevalence of a positive ambulatory ECG response was particularly high (~60 % ) among patients in whom the onset of ST segment depression occurred at a low work load (<75 % of maximal predicted heart rate) (p < 0.001 vs higher work loads) and in those who had a relatively prolonged duration of ST segment depression (~3 minutes) after exercise (p < 0.001 vs lesser dura- tions). (Two patients had missing data for ST duration.) The prevalence of positive ambulatory ECG responses also increased with increasing magnitude of ST depression on exercise testing (p < 0.01).

In the subgroup of patients undergoing 48-hour ambulatory ECG monitoring, there was no significant difference in the frequency of positive ambula-

Volume 125, Number 5, Pan 1 American Heart Journal Klein et al. 1251

Table I. Prevalence of ST-segment depression during 24-hour ambulatory ECG monitoring

24 hr monitoring

FREQUENCY OF POSITIVE AMBULATORY ECG

loo EX ST ONSET t%LYPHR)

60

No. of % (No.) patients positive

%

All patients Exercise ECG

Negative ECG Positive ECG

CAD probability 280 SO CAD probability 21% -79% CAD probability <20% Typical angina

Angiographic CAD Tl scintigraphy:

No reversible defects

244 23 (55)

82 2 (2) 162 33 (53) 119 33 (39) 26 42 (11) 17 18 (3) 32 41 (15)

54 31 (17)

124 15 (19)

%

60

40

20

MPHR>BSX MMR> 75-85x YPHR < 75% ln=333) (n=36) h=30)

loo EX ST MASNITUDE

80

60

40

20

Reversible defects 120 30 (36) Positive exercise ECG 93 39 (36) Negative exercise ECG 27 0 (0)

I-2mm H/D >2-SmmH/D ond > I.5 Up

33mmHlD

h-45) h=35) (n=39) 100 EX ST DURATION

80

tory responses for 24- versus 48-hour monitoring: the frequency in patients with positive exercise responses (n = 47 patients) was 42% for 24-hour monitoring versus 49% for 48-hour monitoring (p not significant). Of 11 patients who had positive results on only 1 day of monitoring, just three (27%) were positive on the second day of monitoring.

Correlation between ambulatory ECG and exercise ECG responses. For subsequent analyses we divided our patients into three groups on the basis of the results of exercise and 24-hour ambulatory ECG monitoring. For these analyses we excluded the patients whose posttest CAD probability after exercise ECG was less than 80 % (n = 87), to confine our analyses to a CAD population (n = 157) consisting of the patients with either a CAD probability of at least 80% or patients with known CAD (i.e., those who had had infarction, coronary artery bypass grafting, or percu- taneous transluminal coronary angioplasty). Group 1 consisted of the 39 CAD patients with both a positive exercise ECG and positive ambulatory ECG response, group 2 consisted of the 80 patients with a positive exercise ECG and negative ambulatory ECG response, and group 3 consisted of the 38 patients with both negative exercise ECG and negative ambulatory ECG responses. (Among this CAD population the fourth combination of a negative exercise ECG and positive ambulatory ECG response was not observed.) The clinical, hemodynamic, and exercise ECG variables in the three patient groups are listed in Table II. Compared with groups 2 and 3, group 1 patients exercised for fewer minutes (p = 0.06) to a lower peak heart rate (p = 0.07) and achieved a lower

% 60

40

20

O-I min >I-bmin *3min - (n=3) (n=71) k43)

Fig. 3. Frequency of positive ambulatory ECG response (vertical axis) was evaluated according to three different ECG criteria. Top, Frequency was assessed in patients according to heart rate at onset of ST-segment depression, expressed as percentage of maximal predicted heart rate (MPHR). Subgroup with lowest threshold for onset of ST depression had highest frequency of positive ambulatory ECG responses (JI < 0.001). Center, Frequency was evaluated when patients were subdivided according to magnitude of ST-segment depression during exercise. The greater the magnitude of ST-segment depression, the higher the frequency of positive ambulatory ECG response (p < 0.01). Bottom, Patients were divided on basis of duration of ST- segment depression into recovery. The more prolonged the duration of ST-segment depression into recovery, the higher the frequency of positive ambulatory ECG responses (p < 0.001). EX, Exercise; HID, horizontal/ downsloping; %MpHR, percentage of maximal predicted heart rate, ST, ST-segment depression; UP, upsloping ST-segment depression.

peak systolic blood pressure (p = 0.05). This was ac- companied by a significantly worse ECG response during exercise testing in the group 1 versus group 2 patients, including a greater magnitude of ST segment depression (p = 0.008), occurring at a lower heart rate threshold (p = 0.001) and persisting for a longer duration after exercise (p = 0.001). Thus group 1 patients had both poorer clinical and exercise ECG responses to exercise.

Thallium 201 stress redistribution scintigraphy. Fig. 4

1252 Klein et al May 1993


Table II. Clinical, hemodynamic, and exercise ECG responses _--__ ._

Group I * (n = XII

___. .-- Clinical and hemodynamic variables

Exercise duration (min) 6.4 t 2.7 Peak exercise HR 138 i 22 Peak exercise SBP 170 i 29 Peak exercise DBP 83 I 11 Exercise chest pain (%) 11 (30)

Exercise ST-segment depression

Magnitude (mm) 3.0 t 0 HR at onset (beats/min) 122 r 19 Min of onset 4.1 +- 2.2 Postexercise duration (min) 3.5 * 2.0

.__-..- ~- - Group -‘I (;mup Xl- tn = 80) ,I = .Ylil /I ctrlur

-- ____-.---

7.5 i- 2.n 7.X -1 .i 0.059 146 + IX 145 1 20 0.068 180 + “4 184 2 :31 0.050

85 t l-1 86 r 13 0.630 18 (231 *i (14) 0.240

2.4 f 0.9 0.008 133 + 18 0.002 5.6 + 2.6 0.001 1.8 + 1.6 0.00 1

Values expressed as mean + SD. DEP, Diastolic blood pressure; HR, heart rate; SBP, systolic blood pressure

*Positive ambulatory ECG; positive exercise ECG. iNegative ambulatory ECG; positive exercise ECG. $Negative ambulatory ECG; negative exercise ECG.

demonstrates the frequency of a positive thallium response in the same three patient subgroups and the number of reversible thallium defects, summed thallium reversibility score, and frequency of TID. Group 1 patients tended to have a greater frequency of positive thallium results compared with the other two groups (p = 0.07), and they manifested a greater median number of reversible thallium defects (p < 0.05) and a greater median summed thallium reversibility score (p < 0.05). The frequency of TID of the left ventricle after exercise was also much higher (p < 0.01) in the group 1 patients. The frequency of abnormal thallium lung uptake was 14 % in group 1, 4% in group 2, and 16% in group 3 (p < 0.05). Thus group 1 patients manifested both a higher frequency and greater magnitude of exercise-induced thallium perfusion abnormalities.

DISCUSSION

Stress redistribution thallium 201 myocardial scin- tigraphy2m4 and ambulatory ECG5-8 both add incre- mental information to exercise ECG in the prognostic assessment of patients with CAD. Nevertheless, the usefulness of ambulatory ECG relative to noninvasive stress imaging modalities remains unclear. First, the prevalence of ST-segment depression during ambulatory ECG is highly debated. Second, there has been relatively little direct comparison of this technique to radionuclide stress testing, the other widely used supplement to exercise ECG. In this study, therefore, we employed stress redistribution thallium 201 myocardial perfusion scintigraphy as an independent standard of myocardial ischemia, to assess the functional significance of ST segment depression during ambulatory ECG monitoring.

The frequency and magnitude of thallium abnor- Prevalence of ST-segment depression during ambu- malities were further assessed in just the 108 patients latory ECG. A positive ambulatory ECG response oc- who had clinically confirmed CAD, on the basis of curred in one third of patients with a positive exer- prior cardiac events (myocardial infarction, bypass cise ECG response in this study. In prior reported surgery, or angioplasty) (Table III). Exclusion of the studies the observed prevalence of ST segment de- high-likelihood CAD patients yielded the same pat- pression during ambulatory ECG monitoring varied tern of results among the remaining patients with markedly, from a low of 23 CA lg to a high of 82 % ,“‘I confirmed CAD: patients with known disease who with an average prevalence of 464‘0 among 1382 had a positive ambulatory ECG response had greater patients accumulated from 10 studies in the litera- values for frequency of positive thallium responses ture.5‘“, lgW2* Differences in patient selection may (p = 0.02), median number of reversible thallium de- have contributed to these wide differences in re- fects (p = 0.02), and median summed thallium reported prevalence, because selection criteria varied versibility score (p = 0.01). This subgroup also had a markedly among these studies. Analysis of subgroups tendency toward a greater frequency of TID (p not in our own patient population supports this explana- significant). tion. For instance, among patients with positive ex-



Table 111. Results of stress redistribution myocardial perfusion scintigraphy in patients with known CAD

Group 1 Group 2 Group 3 P

(n = 20) [n = 55) (n = 33) value

‘% positive Tl result 80 (16) 75 (41) 48(16) 0.02

(No.)

No. of reversible Tl 5.5 (O-13) 4 (O-13) 1 (O-9) 0.02

segments* Summed Tl 9.5 (O-29) 7 (O-27) 3 (O-16) 0.01

reversibility score* % with TID (No.) 30 (6) 11 (6) 15 (5) 0.16

*Data expressed as median (range).

ercise ECG findings who have typical angina-a common selection criterion-the frequency of a positive ambulatory ECG response increased to nearly 50 % . Among positive exercise ECG subgroups, patients with the onset of ST-segment depression at a very low level of exercise, those with prolonged ST- segment depression after exercise and those with increased magnitude of ST depression also had a particularly high frequency of ST-segment depression during ambulatory ECG monitoring.

Increasing the monitoring period to 48 hours did not significantly increase the frequency of a positive ambulatory ECG response. This was because only three of the 23 patients with a positive ambulatory response during 48-hour monitoring had a positive response only on the second day of monitoring. Sim- ilar results were reported by Tzivoni et a1.25 Further evaluation in more patients is needed to determine whether this observation is a consistent finding.

If the exercise ECG response was normal, the likelihood of a positive ambulatory ECG response was extremely low (2%). This confirms the results of other investigations.16,26,27 Thus there is little usefulness in performing ambulatory ECG among patients known to have a negative exercise ECG response.

Functional significance of ambulatory ECG responses: Comparison with exercise ECG. Compared with patients with a negative ambulatory ECG response, the CAD patients with a positive ambulatory ECG response had a significantly greater magnitude, earlier onset, and longer duration of exercise-induced ST- segment depression. Further, they tended to have poorer exercise tolerance and achieved lower peak heart rates and peak systolic blood pressure eleva- tions during maximal exercise. These results suggest that CAD patients with a positive ambulatory ECG

Klein et al. 1253

P-O.07

1” 16

P<O.O5

14. 0 12.

lo-

6.

e-

4. 3 2. 1.5 0 -

C. P<O.O5 26- 24- 20- 16 - - 12 -

6-

4-

0

P<O.Ol

10

0 QFOUP 1 GROUP 2 GROUP 3

(+Amb, +I%) Wnb, +W 6-h -4

Fig. 4. Thallium findings in groups 1, 2, and 3. A, Fre- quency of positive thallium (Tl) response in each group. Group 1 patients had highest frequency of positive thallium response, and group 3 had lowest (p = 0.07). B, Me- dian number and range of reversible thallium defects. C, Median and range for thallium reversibility scores in each group. In B and C data are presented according to “box plot” method. Bottom and top edges of box represent sample’s 25th and 75th percentiles, respectively. Central dark horizontal line is drawn at sample’s median. Central vertical lines extend from box to distance of at most 1.5 times the middle 50% interquartile ranges. Any value more ex- treme than this is marked with a circle. Group 1 patients had greatest median number of reversible segments-five per patient-and highest median thallium reversibility score (p < 0.05). D, shows percentage of patients in each group who demonstrated TID of left ventricle. Group 1 patients had markedly higher frequency of TID compared with groups 2 and 3 (p < 0.01). Amb, Ambulatory ECG; Ex, exercise; Tl, thallium.)

1254 Klein et at

B-

Fig. 5.

PRE-TEST LIKELIHOOD

Test efficacy is dependent on population to which it is applied. This figure illustrates a Bayesian curve for application of exercise ECG. Relationship between pretest probability of CAD (horizontal axis) and posttest probability of CAD (vertical axis) for positive exercise ECG (top, conuex curue) and negative exercise ECG (bottom, concaue curve) is shown. Application of exercise ECG is shown for two hypothetical patients (A and B). Patient B has a very low (5%) pretest probability of CAD. If exercise ECG in such a patient is negative, posttest probability of CAD remains very low (point B- on vertical axis), but if exercise ECG is positive, posttest probability of CAD still remains relatively low (<2041; point B+ on the vertical axis). Thus in asymptomatic adults with low pretest CAD probability there is a greater than four in fifth chance that positive exercise ECG will be false positive. The converse is true if test is applied to a patient with high pretest probability of disease (such as patient A, who has >90% probability of CAD). In such a patient, whereas positive exercise ECG now has high predictive accuracy (point A+ on vertical axis), negative exercise ECG is frequently false negative for CAD (point A-) in more than four fifths of cases. Whereas these curves refer specifically to issue of anatomic diagnosis, they indirectly reflect principles for detecting ischemia per se, as discussed in text.

response are functionally sicker than their counter- parts with a negative ambulatory ECG response.

Functional significance of ambulatory ECG response: Comparison with myocardial perfusion scintigraphy. To confirm the possibility that patients with positive ambulatory ECG responses are functionally sicker, we evaluated the presence and magnitude of exercise- induced hypoperfusion, as assessed by thallium scintigraphy, in our patient subgroups. The CAD patients with positive ambulatory ECG responses manifested a greater frequency of ischemic thallium responses and a greater magnitude of exercise-in-

May 1993


duced abnormalities on planar and SPECT thallium scintigraphy. Specifically, these patients demon- strate a significantly greater median number of reversible thallium defects and greater median summed thallium reversibility score wit,h SPECT imaging. The reversibility score combines the assessments that, measure the extent of inducible hypoperfusion (number of thallium defects) and the severity of inducible hypoperfusion (magnitude of defects). The patients having a positive ambulatory ECG response also demonstrated a striking frequency of TID after stress on the anterior view planar images. One third of the patients with a posit.ive ambulatory ECG had this finding. This planar index is notable because its presence is generally reflective of the presence of both extensive and severe coronary disease, extensive disease being necessary to cause left ventricular dilation and severe disease being necessary to maintain it.l” Thus the planar and SPECT thallium findings were con- cordant with the hemodynamic findings and the results of exercise ECG and confirmed that patients having a positive ambulatory ECG response are a functionally much sicker cohort compared with their coun- terparts with a negative ambulatory ECG response.

False-negative ambulatory ECG responses. The presence of normal exercise and ambulatory ECG responses, however, did not exclude the possibility of myocardial perfusion abnormalities in our CAD patients. For instance, 50’; of group 3 patients (negative exercise and ambulat,ory ECG) had a positive thallium response to exercise, and 16’,! of these patients still demonstrated evidence of TID during thallium scintigraphy (Fig. 4). This observation could be predicted from Bayesian probability analysis, whereby a test’s efficacy in excluding disease de- pends on the underlying prevalence or probability of CAD.g-“.“8 When CAD prevalence is low, the false- positive rate of any noninvasive cardiologic test is at its maximum and the false-negative rate is at its minimum (Fig. 5). Conversely, when the prevalence of CAD is high, the false-positive rate is at, its minimum but the false-negative rate is now maximal. The practical clinical significance of this relationship as it relates to prognosis has been demonstrated in follow-up studies. For instance, Gibson et al.sg found that a positive predischarge exercise ECG had a high predictive accuracy for subsequent cardiac events in patient,s after myocardial infarction; they also observed, however. t,hat 27”,’ of patients without exercise-induced ST-segment depression or angina also had a subsequent event during a mean follow-up period of only 15 -+ 12 months.sg Most of these false- negative ECG responders had thallium evidence of ischemia.

Volume 125, Number 5, Part 1 American Heart Journal Klein et al. 1255

CARDIAC RISK STRATIFICATION

4

r + LOW El RISK

Fig. 6. Algorithm for selection of further tests among patients who have “intermediate” probability of cardiac events after performance of exercise ECG. This algorithm refers only to subgroup of patients in whom exercise ECG can be applied as initial first test (i.e., it is interpretable). If patient has low probability of cardiac events after exercise ECG, then patient may be followed without further testing. If patient has high probability of cardiac events after testing, then aggressive therapy is indicated. In group of patients who have intermediate likelihood of cardiac events after exercise ECG, however, clinical decision making is more challenging and further prognostic testing may be useful. Among such patients radionu- elide testing is indicated for the subgroups who have negative, equivocal, or only mildly positive exercise ECG, because, according to our results, of the relative insensitivity of ambulatory ECG for excluding myocardial ischemia in such instances. In patients with more than mild exercise ECG positivity, both ambulatory ECG and radionuclide imaging may be considered for risk stratification, but negative ambulatory ECG response may not exclude absence of inducible hypoperfusion (i.e., jeopardized myocardium) in this cohort, according to our results.

Similarly, we followed patients with typical angina and a negative exercise ECG.30 We found that a negative exercise ECG had a high false-negative rate for determining ischemia in such patients with high disease probability: 25 % had reversible thallium defects in our cohort. Notably, the l-year cardiac event rate was three times higher in the subgroup with positive thallium responses compared with the subgroup with negative thallium responses, validating the frequent false-negative nature of these normal ECG responses in patients with typical angina.

Comparison with previous studies. Quyyumi et a1.31 recently examined 155 CAD patients undergoing both exercise radionuclide ventriculography and ambulatory ECG monitoring. There are important sim- ilarities between their findings and our own. First, they also found that a negative ambulatory ECG response is frequently false negative for radionuclide indices for ischemia: the ambulatory ECG response was negative in 45 % of patients with an abnormal left ventricular ejection fraction (LVEF) response to ex-

ercise and in one third of patients with up to 10 % fall in LVEF during exercise. However, a positive ambulatory ECG response was associated with more scin- tigraphic evidence of ischemia than a negative response: the mean fall in peak LVEF with exercise was significantly greater (5.6 % t 0.8% ) in patients with a positive ambulatory ECG response compared with those with a negative ambulatory ECG (2.4% ? 0.9 % ). Differences in patient selection, however, limit further comparisons between the study of Quyyumi et al. and our own study.

Implications for test utilization. By using thallium scintigraphy as an independent marker of ischemia, we demonstrated the heightened functional significance of transient ST-segment depression during daily life activity among CAD patients: such patients have both a greater frequency and magnitude of thallium abnormalities during exercise. These results help explain why cardiac event rates are elevated among CAD patients with ambulatory ischemia.5-s Consistent with Bayesian theory, however, a negative

1256 Klein et al.

ambulatory ECG response in CAD patients is frequently false negative for ischemia, as manifested by the relatively high frequency of thallium abnormalities in such patients. This finding appears to limit the clinical usefulness of ambulatory ECG monitoring. Any potential algorithm for using ambulatory ECG monitoring is applicable only for patients with an interpretable exercise ECG, because radionuclide stress testing is clearly the first choice of tests when the exercise ECG cannot be interpreted (e.g., left ventricular hypertrophy, left bundle branch block pat,tern). Sequential prognostic testing is indicated among t,he subgroup of patients who have an intermediate likelihood of cardiac event after an exercise ECG test. Generally, a wide variety of clinical patients fall into such an intermediate likelihood group, ranging from asymptomatic patients with positive exercise ECG responses to patients with typical angina and negative exercise ECG responses.lO, 28 A potential algorithm for choosing between ambulatory ECG and radionuclide stress testing in such intermediate-likelihood patients is depicted in Fig. 6. Ra- dionuclide stress testing is preferable to ambulatory ECG among patients who have either a negative, nondiagnostic, or equivocal exercise ECG response given the very low yield of positive ambulatory ECG responses in such patients. It also appears from our results that radionuclide stress testing is preferable to ambulatory ECG among patients with only weakly positive exercise ECG responses (e.g., those with late-onset ST-segment depression and those with ST-segment depression of short duration after exercise). Among the remaining patients with more positive exercise ECG responses, ambulatory ECG appears to provide useful risk stratification information. If the ambulatory ECG response is positive in such patients, it is evident from both our results and clinical follow-up studies5-s that such patients are at high risk and no further testing may be necessary in these patients. However, a low threshold should be present for performing radionuclide testing if the ambulatory ECG response is negative in such patients, because many of these patients may still have substantial but electrocardiographically silent ischemia according to our results. The cost effectiveness of ambulatory ECG monitoring should be evaluated in the content of this circumscribed clinical application.

We are grateful to Lais Haendchen, MD, Hootan Rahimizadeh. and Sheldon Goldman for their technical assistance, and Frances Katz, Mitzi Escuin, and Astrid Gonzalez for their assistance in manuscript preparation.

REFERENCES

1. Ladenheim ML, Pollack BH, Rozanski A, Berman DS. Staniloff HM, Forrester JS, Diamond GA. Extent and severity of myocardial hypoperfusion as orthogonal indices of

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18

my 1993


prognosis in patients with suspected coronary artery disease. J Am Coil Cardiol 1986;7:464-71. Gibson RS, Watson DD. Cradock GB, Crampton Rd, Kaiser DL, Denny MJ, Belier GA. Prediction of cardiac events after uncomplicated myocardial infarction: a prospective stud! comparing predischarge exercise thallium-201 scintigraph\ and coronary angiography. Circulation 1983;68:321-36. Kaui S, Finkelstein DM, Homma S, Lea&t M, Okada HD, Boucher CA. Superiority of quantitative exercise thallium-20 1 variables in determining long-t,erm prognosis in ambulatory patients with chest pain: a comparison with cardiac catheter- ization. J Am Co11 Cardiol 1988;12:25-34. Kaul S, Lilly DR, Gasho JA, Watson DD, Gibson RS, Oliner CA, Ryan .JM, Belier GA. Prognostic utility of the exercise thallium-201 test in ambulatory patients with chest pain. Cir- culation 1988;77:745-8. Rocco MB, Nabel EG, Campbell S. Goldman L, Barry J, Mead K, Selwyn AP. Prognostic importance of myocardial ischemia detected by ambulatory monitoring in patients with stable coronary artery disease. Circulation 1988;78:877-84. Tzivoni D, Gavish A, Zin D, et al. Prognostic significance of ischemic episodes in patients with previous myocardial infarction. Am-J Cardiol -1988;62:661-4: Deedwania PC, Carbajal EV. Silent ischemia during daily life is an independent predictor of mortality in stable angina. Cir- culation 1990;81:748-56. Gottlieb SO, Gottlieb SH, Achutf SC, Baumgardner R, Mel- lits ED. Weisfeldt ML. Gerstenblith G. Silent ischemia on Holter monitoring predicts mortality in high-risk postinfarc- tion patients. JAMA 1988;259:1030-5. Rozanski A, Berman DS. Silent myocardial ischemia. I. Pathophysiology, frequency of occurrence, and approaches toward detection. AM HEART J 1987;114:615-26. Rozanski A, Berman DS. Silent myocardial ischemia. II. Prognosis and implications for the clinical assessment of patients wit,h coronary artery disease. AM HEART J 1987;114:627- 38. Berman DS, Rozanski A, Knoebel S. The detection of silent ischemia: cautions and precautions. Circulation 1987;75:101-5. Diamond GA, Forrester JS. Analysis of probability as an aid in the clinical diagnosis of coronary artery disease. N Engl d Med 1979;300:13%8.

_ _

Kiat H, Berman DS, Maddahi .I, Yang LD, Van Train K, Rozanski A, Friedman J. Late reversibility of tomographic myocardial thallium-201 defects: an accurate marker of myocardial viability. J Am Co11 Cardiol 1988;12:1456-63. Weiss AT, Berman DS, Lew AS, et al. Transient ischemic dilation of the left ventricle on stress thallium-201 scintigraphy: a marker of severe and extensive coronary artery disease. J Am Co11 Cardiol 1987;9:752-9. Levy R, Rozanski A, Berman DS, Garcia E, Van Train K. Maddahi d, Swan HJC. Analyses of the degree of pulmonary thallium “washout” following exercise. .J Am Co11 Cardiol 1983;2:719-28. Tzivoni D. Benhorin J, Gavish A, Stern S. Holter recording during treadmill testing in assessing myocardial ischemic chances. Am J Cardiol 1985;55:1200-3. Hed&s SL. Krantz DS, Centroda RJ, Rosen S, Rozanski A. Development of a diary for use with ambulatory monitoring of mood. activities and phvsioloaic function. J Psychopath01 Be- hav Assess 1990;12:203:12. -

.

Lin LI-K. Concordance correlation coefficient to evaluate re- producibility. Biometrics 1989:45:255-68.

19. Von Arnim T, Hofling B, Schreiber M. Characteristics of episodes of ST elevation or ST depression during ambulatory monitoring in patients subsequently undergoing coronary an- eiographv. Br Heart J 1985;54:484-8.

20. ?vIo>vFi’, Nademanee K, Intarachot V, Josephson MA, Rob- ertson HA. Sinah BN. Severity of silent myocardial ischemia

L

on ambulatory electrocardiographic monitoring in patients with stable angina pectoris: relation to prognostic determi- nants during exercise stress testing and coronary angiography. J Am Co11 Cardiol 1988;12:1169-76.

21. Cecchi AN. Dnvellini EV. Marchi F, Puci P. Santoro GM,


American Heart Journal Bach et al.

Fassini PF. Silent myocardial ischemia during ambulatory electrocardiographic monitoring in patients with effort angina. J Am Co11 Cardiol 1983;3:934-9.

22. Mulcahy D, Keegan J, Crean P, Quyyumi A, Shapiro L, Wright C, Fox K. Silent myocardial ischaemia in chronic stable angina: a study of its frequency and characteristics in 150 patients. Br Heart J 1988;60:417-23.

23. Gunther H, Osterspey A, Tris-Muller I, Eggeling T, Hopp HW, Hilger HH. The sensitivity of 24 h Holter monitoring and exercise testing for the recognition of myocardial ischemia: a comparative study. Eur Heart J 1988;(suppl N):46-9.

24. Cohn PF, Ventrovec GW, Nesto R, Gerber FR, Total Ischemia Awareness Program Investigators. The nifedipine-total ischemia awareness program: a nation survey of painful and pain- less myocardial ischemia including results of antiischemic therapy. Am J Cardiol 198963:534-g.

25. Tzivoni D, Gavish A, Benhorin J, Banai S, Keren A, Stern S. Day-to-day variability of myocardial ischemic episodes in coronary artery disease. Am J Cardiol 1987;60:1003-5.

26. Campbell S, Barry J, Rocco MB, et al. Features of the exercise test that reflect the activity of ischemic heart disease out of hospital. Circulation 1986;74:72-80.

27.

28.

29.

30.

31.

Quyyumi A, Crake T, Wright C, Mockus L, Fox K. The role of ambulatory ST-segment monitoring in the diagnosis of coronary artery disease: comparison with exercise testing and thallium scintigraphy. Eur Heart J 1987;8:124-9. Rozanski A, Berman DS. The efficacy of cardiovascular nuclear medicine studies. Semin Nucl Med 1987;27:104-20. Gibson RS, Watson DD, Craddock GB, Crampton RS, Kaiser DL, Denny MJ, Beller L. Prediction of cardiac events after uncomplicated myocardial infarction: a prospective study comparing predischarge exercise thallium-201 scintigraphy and coronary angiography. Circulation 1983;68:321-36. Bairey CN, Rozanski A, Maddahi J, Resser KJ, Berman DS. Exercise thallium-201 scintigraphy and prognosis in patients with typical angina and negative exercise electrocardiography. Am J Cardiol 1989;64:282-7. Quyyumi AA, Panza JA, Diodati JG, Dilsizian V, Fallahan TS, Bonow RO. Relation between left ventricular function at rest and with exercise and silent myocardial ischemia. J Am Co11 Cardiol 1992;19:962-7.

Dobutamine stress echocardiography: Prevalence of a nonischemic response in a low-risk population

The problems of population referral bias in the calculation of specificity in diagnostic testing for coronary artery disease have been previously described. Previous studies investigating the sensitivity and specificity of dobutamine stress echocardiography (DSE) have been subject to pretest and posttest referral biases, largely as a result of the requirement for coronary arteriography. This study determines the normalcy rate for DSE by examining a population at statistically low risk for coronary artery disease. The probability of significant coronary artery disease was determined for 828 consecutive patients referred for DSE at the University of Michigan, and groups were identified with <iO% and <5% probability of disease. Four of 72 patients (5.6%) with a normal baseline echocardiogram and a probability of coronary artery disease of <lo%, and three of 38 (7.9%) with a probability of <5% were found to have an abnormal DSE, yielding normalcy rates of 94.4% and 92.1%, respectively. The area of abnormality involved the posterior circulation in three of four patients (75%). This study demonstrates that DSE has a normalcy rate of 92% to 94% and is an accurate test for excluding the presence of significant coronary artery disease. (AM HEART J 1993;125:1257.)

David S. Bach, MD, Anne Hepner, MD, Pamela A. Marcovitz, MD, and

William F. Armstrong, MD Ann. Arbor, Mich.

Dobutamine stress echocardiography (DSE) is an accurate and feasible method for the noninvasive de-

From the Division of Cardiology, Department of Internal Medicine, University of Michigan Medical Center.

Received for publication Sept. 21, 1992; accepted Nov. 6, 1992.

Reprint requests: David S. Bach, MD, University of Michigan Hospital, UH BlF245-Box 0022, 1500 E. Medical Center Dr., Ann Arbor, MI 48109.

Copyright $ 1993 by Mosby-Year Book, Inc. 0002-8703/93/$1.00 + .lO 4/l/44922

tection of coronary artery disease.l-lo Using selective coronary angiography as a measure of accuracy, the sensitivity of DSE for the detection of coronary artery disease has been high, ranging from 68% to 97 % .3-10 The specificity of DSE, however, has been more variable, ranging from as high as 1OO%4 to as low as 45 % .3 Several factors may be responsible for the low test specificity in these studies. In addition to the possibility of error in test interpretation, pretest

1257

prevalence and functional significance of transient st-segment depression during daily life...

Documents