- Email: [email protected]
The value of silent myocardial ischemia monitoring in the prediction of perioperative myocardial infarction in patients undergoing peripheral vascular surgery
The value of silent myocardial ischemia monitoring in the prediction of perioperative myocardial infarction in patients undergoing peripheral vascular surgery Peter F. Pasternack, M D , Eugene A. Grossi, M D , F. Gregory Baumann, PhD, Thomas S. Riles, M D , Patrick J. Lamparello, M D , Gary Giangola, M D , Lawrence K. Primis, BA, Ronnie Mintzer, BS, and A n t h o n y M. Imparato, M D ,
New York, N.Y. Real-time electrocardiographic monitoring for silent myocardial ischemia was performed on 200 patients undergoing peripheral vascular surgery to try to better define those at high risk of perioperative myocardial infarction. The patients were divided into those undergoing abdominal aortic aneurysm or lower extremity revascularization procedures (group I, n = 120) and those undergoing carotid artery endarterectomy (group II, n = 80). Silent ischemia was detected during the preoperative, intraoperative, or postoperative periods in 60.8% of group I and 67.5% of group II patients. Six group I and three group II patients suffered an acute perioperative myocardial infarction with two cardiac deaths. In both groups I and II a variety of parameters based on monitoring of silent myocardial ischemia were compared between the subgroups of patients who had myocardial infarction and those who did not.The results show that in both groups there was a significantly (p -< 0.05) greater total duration of perioperative ischemic time, total number of perioperative ischemic episodes, and total duration of perioperative ischemic time as a percent of total monitoring time in patients who suffered a perioperative myocardial infarction compared to those who did not. Multivariate logistic regression analysis of preoperative characteristics in all 200 patients showed the occurrence of preoperative silent myocardial ischemia and angina at rest to be the only significant predictors of perioperative myocardial infarction. Thus perioperative monitoring for silent myocardial ischemia might n0ninvasively identify those patients undergoing peripheral vascular surgery who are at increased risk for perioperative myocardial infarction, permitting implementation of timely preventive measures in selected patients. (J VASCSURG 1989;10:61725.)
Coronary artery disease is the most common cause of early postoperative death after abdominal aortic aneurysm resection, 1 lower extremity revascularization, 2 and carotid endarterectomy. 3-s Although coronary angiography can define the extent of coronary artery disease, this diagnostic approach cannot noninvasively predict which patients are most likely to suffer acute myocardial infarction (MI) and cardiac death in association with peripheral Vascular From the departmentsof Medicineand Surgery,New YorkUniversity MedicalCenter. Presented at the ThirteenthAnnualMeeting of the SouthernAssociation for Vascular Surgery, Key West, Fla., Jan 25-28, 1989. Reprint requests: Peter F. Pasternack, MD, Suite 4C, Faculty Practice, New YorkUniversityMedicalCenter, 530 First Ave., New York, NY 10016. 24/6/15572
surgical procedures. Other noninvasive methods, such as resting gated blood pool studies, 6'7 have been tested and have proved of some value in indicating those patients undergoing peripheral vascular surgery who are at the highest risk of a perioperative myocardial infarction, but no previous method has shown optimal accuracy. However, recent technologic advances now permit detection and quantitation of physiologic manifestations of myocardial ischemia in the absence of angina pectoris, a condition termed silent myocardial ischemia. Measurement of such silent myocardial ischemia might prove to be a useful noninvasive predictor of perioperative myocardial infarction in patients undergoing peripheral vascular surgical procedures. Therefore in the present study we tested the ability of a new, highly sensitive device for detecting and 617
618
Journal of VASCULAR SURGERY
Pasternack et al.
quantitating silent myocardial ischemia to distinguish the subgroup of patients at high risk for perioperative myocardial infarction in association with various peripheral vascular surgical procedures.
METHODS Patient population Two hundred patients met the criteria for admission to the study out of a total of 227 consecutive patients who underwent a peripheral vascular surgical procedure at New York University Medical Center from September 1987 to June 1988. Nineteen patients were excluded because of a QRS duration greater than 120 msec and eight patients were excluded because of left ventricular hypertrophy with marked repolarization changes. The study group included 157 men and 43 women with a mean ± SD age of 69.5 _ 8.37 years (range, 48 to 89 years). These patients had the following incidences of clinical evidence of coronary artery disease: history of exertional angina in 32%, history of rest angina in 9%, history of prior MI in 26%, and electrocardiographic (ECG) evidence of a prior MI in 14%. The 200 patients included in the study underwent one of three types of peripheral vascular surgical procedures. Sixty-seven patients (33.5%) underwent abdominal aortic aneurysm repair, 53 patients (26.5%) had lower extremity revascularization procedures, and 80 patients (40.0%) underwent carotid endarterectomy. The degree of operative stress, an important factor in the incidence of myocardial injury, differs considerably among these types of operative procedures, especially inasmuch as 75% (60/80) of the carotid endarterectomy procedures were performed with the patient under local anesthesia. Therefore the patients were subdivided into two main groups for the purposes of analysis. Group I comprised all patients undergoing abdominal aortic aneurysm or lower extremity revascularization procedures (n = 120, 60.0%). Group II contained all patients who underwent carotid endarterectomy procedures (n = 80, 40.0%). The 120 patients in group I had a mean + SD age of 70.0 _+ 8.54 years (range, 48 to 89 years) and 103 (85.8%) were men. These patients exhibited the following incidences of clinical evidence of coronary artery disease: history of exertional angina in 31.7%, history of rest angina in 7.5%, history of prior MI in 26.7%, and ECG evidence of a prior MI in 16.7%. The 80 patients in group II had a mean ± SD age of 68.9 + 8.11 years (range, 48 to 83 years) and 54 (67.5%) were men. This group manifested the following incidences of clinical evidence of coronary
artery disease: history of exertional angina in 32.5%, history of rest angina in 11.2%, history of prior MI in 25.0%, and ECG evidence of a prior MI in 11.2%.
Digital electrocardiographic monitor Patients were monitored for silent myocardial ischemia by means of a real-time ambulatory device, the QMED Monitor One Star (QMed Inc., Clark, N.J.), which continuously records and analyzes the electrocardiogram.8,9This monitor acquires the ECG signal through rigid, homogeneous, nonanodized, silver-silver chloride electrodes designed for longterm monitoring. The incoming signal is digitized at 256 Hz and passed to a 65CO2 microprocessor for analysis. Signal amplitude is compared with a fixed internal reference used for calibration. Frequency response is uniform from 0.05 to 40 Hz. The signal is analyzed by an algorithm, which is segmented into a low-level beat detection segment and a high-level analysis segment. The algorithm detects and validates a QRS complex as a region of sustained slope that meets criteria for duration, peak amplitude, and number of inflection points. Validated complexes are then analyzed to determine the onset, peak J-point, and J2 point (60 msec after J-point). The algorithm adjusts continuously to changing width of the QRS complex, and in the default condition it will not measure ST deviation in the presence of a QRS 120 msec or longer. Rapid slope changes (more than 0.15 mV in 60 msec) during the ST segment (J1 to J2) cause the device to exclude the beat from analysis for ST deviation. An instantaneous ST segment amplitude is measured at the J2 point by comparison to the amplitude of the preceding PR interval and is corrected for baseline variation. The instantaneous amplitude is used to update the ST level average on a beat-bybeat basis, and the average is stored in a series of shift registers. The definition of the amplitude of an ischemic event is selectable at - 0.1, - 0.2, - 0.3, or - 0 . 4 mV. The moment when the ST segment descends to - 0 . 1 mV is marked as the potential beginning of an ischemic event. If depression is sustained for approximately 40 seconds, the algorithm validates the ischemic event as real and begins measuring its duration. Thus time is used as a filter for accurately defining significant ST deviation. The duration of an individual depression from the point of validation to a point approximately 40 seconds after the return of the ST segment above - 0 . 1 mV is noted. The ischemic event is also segmented into durations of ST depression at - 0.1, - 0.2, - 0.3,
Volume 10 Number 6 December 1989
- 0.4 mV, and the durations of all events and segments at these levels are summed and averaged for the monitoring period. Quantitative and waveform data are stored in solid-state, random-access memory for subsequent retrieval via a liquid crystal display or a printer. A real-time clock permits the printing of time of day with each cardiac event strip. Perioperafive monitoring During the perioperative period, QMED electrodes were attached in a lead CMs configuration, with the negative reference over the sternum and positive electrode in the Vs position. The earth electrode was attached over the right lower anterior ribs. The electrodes were then connected to the recorder, and monitoring was started during the preoperative period (mean _+ SD preoperative monitoring interval: group I, 18.3 _+ 17.0 hours; group II, 1 8 . 7 _ 16.9 hours; p > 0.05) and continued throughout the operation (mean -+ SD intraoperative monitoring interval: group I, 5.46 _+ 1.92 hours, group II 3.57 -+ 1.15 hours;p < 0.001) and postoperative periods (mean --- SD postoperative monitoring interval: group I, 34.1 _+ 19.6 hours; group II, 32.4 _ 13.3 hours; p > 0.05). An ischemic ST segment change was defined as 1 mm or more of planar or downsloping ST segment depression in lead CMs. The output of the monitor consisted of indications of the presence or absence ofischemic changes and the duration and magnitude of any ST segment depression. The monitor's algorithm excludes ectopic beats and noise and identifies isoelectric, J, and J + 60 msec points in the beats classified as normal. Downsloping or horizontal ST depression of at least 1 mm leasting more than 40 seconds represents the threshold for detection of an ischemic event, but timing of event duration does not commence until the end of the 40 second validation period. The event is considered to be over when the ST amplitude has been above this 1 mm threshold for longer than 40 seconds. For each patient, up to 91 short strips of ECG representative of control and ischemic episodes were available for visual confirmation as part of the printed output. The diagnosis of acute perioperative MI was made on the basis of abnormal elevation of the serum creatine phosphokinase (CPK) MB isoenzyme level as determined daily for the first 5 postoperative days. Data collection and statistical analysis The study provided for the collection of 91 data points on each patient related to general clinical in-
Silent myocardial ischemia monitoring 619
formation such as medical history and results of diagnostic tests or to overall monitoring information. These data were obtained by reviewing the patient's hospital record or by direct questioning of the patient and by examining the record from the special ECG monitor. In addition, for each episode of silent myocardial ischemia for each patient 14 additional monitoring data points related to the ischemia, such as the time of day when the ischemia began and the duration of the episode, were gathered from the printout from the special ECG monitor. All these data were entered into a computer and analyzed by means of the statistical software packages SPSSx (SPSS Inc., Chicago, Ill.) or SAS (SAS Institute Inc., Cary, N.C.). All continuous data are expressed as mean + SD. For continuous variables the significance of differences between groups I and II was determined by the Student's t test. Within group I or II the significance of differences in continuous variables between patients who suffered an MI and patients who avoided MI was determined by a nonparametric method, the Mann-Whitney U test (Wilcoxon rank sum W test). For categoric variables the significance of any differences between the group that suffered a perioperative MI and the group that avoided an MI was analyzed by the chi-square test. Logistic regression analysis was used for multivariate analysis of categoric variables to determine the best set of variables independently predictive of perioperative MI. The estimate of the coefficient in the regression equation, the estimate of the odds ratio, and the 95% confidence interval for the odds ratio were maximum likelihood estimates. The odds ratio is the relative odds of the occurrence of a perioperative MI associated with a change in the predictor variable. RESULTS Silent ischemia Overall, the occurrence of one or more episodes of silent myocardial ischemia was detected in 127 patients (63.5%) during the preoperative, intraoperative, or postoperative periods. In group I, one or more episodes of silent myocardial ischemia appeared in 73 patients (61%) during one or more of these periods. During a period of preoperative monitoring of 21.2 + 19.3 hours, 48 patients (40%) showed evidence of silent ischemia. Among these patients the number of preoperative ischemic episodes was 13.0 + 16.7, the total duration of preoperative ischemia was 5.4 + 8.2 hours, and the total duration of preoperative silent ischemia as a per-
Journal of VASCULAR
620 Pasternack et aI.
SURGERY
Table I. Characteristics of silent myocardial ischemia among 120 patients with abdominal aortic aneurysm or lower extremity revascularization (group I) who suffered myocardial ischemia~ Preoperative Intraoperative Postoperative No. of patients with ischemia (%) Duration of monitorhag (hr) Duration ofischemia (hr) No. ofischemic episodes Total duration of silent myocardial ischemia as % of monitoring interval
48 (40.0%)
45 (37,5%)
57 (47.5%)
21.2 -+ 19.3
5.40 _+ 1.60
38.6 + 19.2
5.37 + 8.18
1.58 + 2.38
7.13 -+ 11.6
13.0 + 16.7
4.44 _+_4.10
12.7 + 13.7
29.0 + 36.2%
32.4 _+ 49.6%
20.5 _+ 30.9%
~Patients had myocardial ischemia during the preoperative, intraoperative, or postoperative periods (mean _+ SD).
Table II. Characteristics of silent myocardial ischemia among 80 patients with carotid artery endarterectomy (group II) who suffered myocardial ischemia~ No. of patients with ischemia (%) Duration of monitoring (hr) Duration ofischemia (hr) No. ofischemic episodes Total duration of silent myocardial ischemia as % of monitoring interval
Preoperative
Intraoperative
Postoperative
30 (37.5%)
33 (41.2%)
43 (53.8%)
22.9 _+ 21.1
3.82 -+ 1.61
32.9 + 11.8
5.16 _+ 11.0
0.65 + 0.85
4.74 +_ 8.47
15.4 _+ 19.3
3.36 _+ 2.74
17,6 _+ 19.4
21.9 +_ 30.8%
19.7 _+ 26.4%
14.3 + 24.8%
*Patients had myocardial ischemia during preoperative, intraoperative, or postoperative periods (mean _+ SD).
centage of the total duration of preoperative monitoring time (percent time ischemic) was 29.0% _+ 36.2% (Table I). During a period of intraoperative monitoring of 5.4 _+ 1.6 hours, 45 patients (38%) showed evidence of silent ischemia. In these patients the number of intraoperative ischemia episodes was 4.4 _+ 4.1, the total duration of intraoperative ischemia was 1.6 _+ 2.4 hours, and the percent time ischemic was 32.4% _+ 49.6%. During a period of postoperative monitoring of 38.6 _+ 19.2 hours, 57 patients (48%) showed evidence of silent ischemia. Among these patients the number of postoperative ischemia episodes was 12.7 _+ 13.7, the total duration of postoperative ischemia was 7.1 _+ 11.6 hours, and the percent time ischemic was 20.5% _+ 30.9%. In group II, one or more episodes of silent myocardial ischemia were found in 54 patients (68%)
during the preoperative, intraoperative, or postoperative periods. During a period of preoperative monitoring of 22.9 _+ 21.1 hours, 30 patients (38%) had evidence of silent ischemia. Among these patients the number of preoperative ischemic episodes was 15.4 _+ 19.3, the total duration of preoperative ischemia was 5.2 + 11.0 hours, and the percent time ischemic was 21.9% + 30.8% (Table II). During a period of intraoperative monitoring of 3.8 _+ 1.6 hours, 33 patients (41%) showed evidence of silent ischemia. In these patients the number of intraoperative ischemic episodes was 3.4 _+ 2.7, the total duration of intraoperative ischemia was 0.65 _ 0.85 hours, and the percent time ischemia was 1 9 . 7 % +_ 26.4%. During a period of postoperative monitoring of 32.9 _ 11.8 hours, 43 patients (54%) showed evidence of silent ischemia. Among these patients the number of postoperative ischemic episodes was
Volume 10 Number 6 December 1989
Silent myocardial ischemia monitoring
621
Table III. Indexes of silent ischemia among 120 patients with abdominal aortic aneurysms or lower extremity revascularization (group I) + Myocardial infarction No myocardialinfarction p value No. o f patients Time monitored preoperatively (hr) Time monitored intraoperatively (hr) Time monitored postoperatively (hr) Total time monitored perioperatively (hr) Duration o f preoperative SI (hr) Duration o f intraoperative SI (hr) Duration ofpostoperatire SI (hr) Total duration ofperioperative SI (hr) No. o f preoperative SI episodes No. ofintraoperative SI episodes N o o f postoperative SI episodes Total no. ofperioperative SI episodes Duration o f SI as % o f preoperative monitoring interval Duration o f SI as % o f intraoperative monitoring interval Duration o f SI as % o f postoperative monitoring interval Total duration o f SI as % o f total perioperative monitoring interval
6 21.8 +- 21.0
114 18.2 -- 16.9
-NS
5.78 -4- 2.36
5.44 _+ 1.90
NS
46.8 z 19,6
33.4 -+ 19.4
NS
74.4 -4- 34.0
56.7 _+ 25.2
NS
3.73 _+ 4.07
2.06 -+ 5.86
0.036
2,40 + 4.05
0.50 -- 1.39
NS
11.5 + 13.1
2.96 _+ 8.32
0.001
I7.7-+ 17.8
5.52 + 13.5
0.002
11.5 + 19.0
5.49 +_ 11.6
0.014
4.50 -~ 5.05
2.18 + 2.84
NS
26.2 + 17.2
5.54 +_ 9.74
0.001
42.2 + 25,4
13,2 + 18.1
0,001
31.2 -+ 41.6%
10.6 +_ 25.7%
0,035
30.4 -+ 44.1%
11.2 + 33.4%
NS
23.8 -+ 25.4%
9.02 + 23.3%
0.001
24.0 +_ 26.3%
9.84 -+ 23.8%
0.003
NS, Not significant; SI, silent myocardial ischemia; mean + SD. +Includes patients with myocardial infarction and those who avoided myocardial infarction.
17.6 + 19.4, the total duration of postoperative ischemia was 4.7 ___ 8.5 hours, and the percent time ischemic was 14.3% _+ 24.8%. Myocardial infarction Six patients in group I (three patients with abdominal aortic aneurysm and three patients with' lower extremity revascularization) and three in group II suffered acute perioperative MIs with CPK-MB levels present in the serum within the first 48 hours after surgery. All nine patients displayed evidence of silent ischemia during the preoperative, intraoperative, or postoperative periods, and there were no perioperative MIs among patients who escaped perioperative silent myocardial ischemia. There were two cardiac deaths among the 200 patients, both in patients in group I who had shown
evidence of silent myocardial ischemia. One suffered an extension of non-Q wave MI and died on the twelfth postoperative day. The other patient, who did not suffer a perioperative MI but who had evidence of preoperative, intraoperative, and postoperative ischemia, had refractory congestive heart failure and died in ventricular fibrillation on the twelfth postoperative day. In group I there was no significant difference in preoperative, intraoperative, postoperative, or total perioperative time monitored between those patients who suffered a perioperative MI and those with no MI (Table III). However, the total duration of pcrioperative silent ischemia, the total number of perioperative episodes of silent ischemia, and the total duration of perioperativc silent ischemia as a percentage of total perioperative time monitored were
622
Journal of VASCULAR SURGERY
P a s t e r n a c k e t al.
T a b l e I V . Indexes o f silent ischemia a m o n g 80 patients with carotid artery endarterectomy (group II) *
No. of patients Time monitored preoperatively (hr) Time monitored intraoperatively (hr) Time monitored postoperatively (hr) Total time monitored perioperatively Total duration ofpreoperative SI (hr) Total duration ofintraoperative SI (hr) Total duration ofpostoperative SI (hr) Total duration of perioperative SI (hr) No. of preoperative SI episodes No. ofintraoperative SI episodes No. of postoperative SI episodes Total no. ofpetioperative episodes Total duration of SI as % of preoperative monitoring interval Total duration of SI as % of intraoperative monitoring interval Total duration of SI as % of postoperative monitoring interval Total duration of SI as % of total perioperative monitoring interval
Myocardial infarction
No myocardial infarction
p value
3 14.0 -+ 1.83
77
--
18.8 +_ 17.2
NS
3.94 -- 0.82
3.55 + 1.16
NS
38.3 _+ 13.1
32.2 -+ 13.3
NS
53.1 --- 18.9
56.2 + 14.6
NS
0.04 + 0.03
2.01 - 7.26
NS
0.42 _+ 0.20
0.26 -+ 0.64
0.033
6.78 _+ 5.68
2.38 +- 6.63
0.016
7.24 + 5.60
4.65 -+ 12.7
0.050
2.00 _+ 1.00
6.57 -+ 13.9
NS
3.33 + 1.53
1.92 + 2.11
0.011
39.3 + 23.5
8.79 + 15.2
0.009
44.7 _+ 23.5
17.3 + 24.0
0.029
0.24 _+ 0.23%
8.51 -+ 21.8%
NS
10.6 _+ 3.89%
8.04 -+ 19.8%
0.037
16.1 _+ 8.15%
7.39 + 19.7%
0.023
11.9 _+ 6.07%
7.43 -+ 18.2%
0.050
NS, Not significant; SI, silent myocardial ischemia; mean _+ SD. *Includes patients with myocardial infarction and those who avoided myocardial infarction.
significantly greater in those patients w h o suffered an acute perioperative M I c o m p a r e d to those with n o MI. T h e patients w h o suffered an acute M I s h o w e d a total duration o f perioperative silent myocardial ischemia o f 17.7 -- 17.8 hours c o m p a r e d to 5.52 _ 13.5 hours for those patients with n o M I (p = 0.002). Likewise, the patients w h o suffered an M I experienced a total o f 42.2 _+ 25.4 episodes
of perioperative silent ischemia compared to a total o f 13.2-_+ 18.1 episodes in those with n o M I (p = 0.001). In addition, the total duration o f p e r i operative silent ischemia as a percentage o f the total perioperative time m o n i t o r e d (percent time ischemic) was significantly higher in those suffering an acute perioperative M I c o m p a r e d to those with no MI. T h e patients w h o demonstrated elevation o f C P K - M B had a perioperative percent time ischemic
o f 24.0% --_ 26.3% c o m p a r e d to 9.84% + 23.8% for those with n o M I (p = 0.003). Similarly, in g r o u p II there was n o significant difference in preoperative, intraoperative, postoperative, or total perioperative time m o n i t o r e d between those patients w h o suffered a perioperative M I and those with n o M I (Table IV). H o w e v e r , the total duration o f perioperative silent ischemia, the total n u m b e r o f perioperative episodes o f silent ischemia, and the total duration ofperioperative silent ischemia as a percentage o f total perioperative time m o n i t o r e d were significantly greater in those patients w h o suffered an acute perioperative M I c o m p a r e d to those with no MI. The patients w h o suffered an acute M I s h o w e d a total duration o f perioperative silent myocardial ischemia o f 7.24 _+ 5.60 hours c o m p a r e d to 4.65 _+ 12.7 hours for those patients with n o M I
Volume 10 Number 6 December 1989
(p < 0.05). Also, the patients who suffered an MI experienced a total of 44.7 _+ 23.5 episodes of perioperative silent ischemia compared to a total of 17.3 + 24.0 episodes in those with no MI (p = 0.03). Finally, the patients who demonstrated elevation of CPK-MB had a perioperative percent time ischemic of 11.9% + 6.07% compared to 7.43% __+ 18.2% for those with no MI (p -< 0.05). Multivariate analysis of categoric perioperative factors was performed with a logistic regression model to determine which factors might be significant predictors of perioperative MI. In addition to the occurrence of preoperative silent myocardial ischemia, the variables tested included a history of a previous MI, hypertension, angina at rest, angina on exertion, congestive heart failure, and EGG evidence of a previous MI. When the patients were divided into the relatively small groups I and II and these were tested separately, group I showed only the presence of angina at rest to be a significant predictor of perioperative MI (beta coefficient in logit model = 2.89, p = 0.002), with an odds ratio of 18.0 (95% confidence interval = 2.97, 109). In group II none of the variables tested proved to be a significant predictor of MI. However, testing of the entire group of 200 patients undergoing any of three types of peripheral vascular surgery showed the presence of preoperative silent myocardial ischemia (beta = 1.56, p = 0.04) and angina at rest (beta = 1.66, p = 0.04) to be significant predictors of perioperative MI with odds ratios of 4.76 (95% confidence interval = 1.03, 22.0) and 5.2 (95% confidence interval = 1.04, 22.6), respectively. DISCUSSION As noted, coronary artery disease remains the leading cause of early and late mortality after peripheral vascular surgical procedures. Acute myocardial infarction is the cause of about one half of all postoperative deaths after abdominal aortic aneurysm repair, 1°12 carotid endarterectomy,saa or lower extremity revasculatization procedures. 12'14 Longterm survival is also substantially limited by a high incidence of late cardiac deaths, especially among those with obvious evidence of coronary artery disease at the time of their initial presentationY -~s Thus an easily accessible, inexpensive, noninvasive method of identifying those patients who are likely to suffer or even die of an MI in association with peripheral vascular surgical procedures would be extremely helpful in perioperative management. The digitized real-time ischemia monitor used in this
Silent myocardial ischemia monitoring 623
study appears to be a promising candidate to fill this role. The rcsuks of this study show that there is a previously unappreciated, extensive amount of silent myocardial ischemia present in patients undergoing peripheral vascular surgical procedures, with 63.5% of 200 patients (60.8% in group I, 67.5% in group II) suffering silent myocardial ischemia during the preoperative, intraoperative, or postoperative period. It is likely that this high incidence of silent ischemia has an anatomic correlation and is a manifestation of severe underlying coronary artery disease in this patient population. In an earlier study cardiac catheterization of 1000 patients undergoing abdominal aortic aneurysm surgery, lower extremity revascularization, or carotid endartcrectomy demonstrated significant coronary artery disease (>70% stenosis) involving a single coronary artery in 27%, two coronary arteries in 19%, and three coronary arteries in 11% of these patients. 16 Recent technologic advances have rapidly accelerated the accumulation of knowledge regarding silent myocardial ischemia. In 1961 Holter 17observed and described an apparent episode of silent ischemia in a patient known to have angina. Changes observed on an ambulatory ECG recording were equivalent to those sccn. during angina but were not associated with pain. Since then the development of accurate monitoring of the ST segment has shown that patients with coronary artery disease have frequent episodes of such painless myocardial ischemia during daily activity. In fact, 75% of all ischemia episodes have been found to be painless. Is Such silent ischemia can occur even at low heart rates and in the absence of physical exercise. Thus silent ischemia is simply myocardial ischemia that occurs in the absence of chest pain and results from the same sequence of physiologic events as ischemia accompanied by angina. Conjecture as to why myocardial ischemia is so often silent has involved several hypotheses including: silent ischemia indicates less severe ischemia of a degree insufficient to reach pain thresholds; there are differences in pain perception thresholds or central transmission of painful stimuli among individuals; and there may be different pathophysiologic characteristics resulting in different patterns of myocardial blood flow distribution during silent, as opposed to symptomatic, ischemia. In addition, there has been much recent speculation regarding the potential prognostic implications, if any, of the common phenomenon of silent
624
Pasternack et al.
ischemia. The data obtained from this study suggest that the frequency and duration of perioperative episodes of silent ischemia are significantly associated with a higher incidence of acute perioperative MI. Three different primary indexes of the severity of silent ischemia were tested for groups I and II in this study: the total duration of perioperative ischemia, the total number of perioperative ischemic episodes, and the total perioperative percent time ischemic. In both groups the total duration of episodes of perioperative silent ischemia was found to be significantly longer in patients suffering an acute perioperative MI compared to those with no MI (Tables III & IV). In addition, in both groups I and II the total number of perioperative ischemic episodes and the total perioperative percent time ischemic were significantly longer in those suffering an acute perioperative MI compared to those with no MI (Tables III & IV). Thus the present findings show that perioperative silent ischemia is of significantly longer total duration and significantly more frequent in those patients who suffer acute perioperative MI. The longer total duration and more numerous episodes of perioperative ischemia in patients who suffer MI may play a role in creating irreversible myocardial damage during the perioperative period. For the total study group of 200 patients multivariate analysis of most available factors that might be considered predictors of perioperative MI, including a history of previous MI, hypertension, angina at rest, angina on exertion, congestive heart failure, ECG evidence of previous MI, and the presence of preoperative silent myocardial ischemia, showed only angina at rest and preoperative silent myocardial ischemia to be significant predictors of perioperative MI. A patient with angina at rest had 4.7 times the risk of a perioperative MI in association with peripheral vascular surgery as a patient without angina at rest. A patient who manifested silent myocardial ischemia preoperatively had 5.2 times the risk of a perioperative MI as a patient without preoperative silent myocardial ischemia. However, it should be noted that since the model used for logistic regression analysis is exploratory in nature, any results regarding the predictive value of certain variables with respect to the occurrence of perioperative MI should be validated by further studies. These findings regarding the predictive value of preoperative silent ischemia in patients who are to undergo peripheral vascular surgery is encouraging with respect to the goal of timely identification of those patients who are more likely to suffer an acute perioperative MI. In this study preoperative moni-
Journal of VASCULAR SURGERY
toring was initiated less than 24 hours before surgery, a lead time that should be of practical value for early identification of patients with significant evidence of preoperative silent ischemia. As we have noted in previous work, 6,7 those patients undergoing peripheral vascular surgery who have low cardiac ejection fractions as determined by preoperative resting gated blood pool studies have a significantly higher incidence of perioperative MI than patients with higher ejection fractions. Similarly analyzed data from the present study of a total of 200 patients showed that the incidence of perioperative MI among patients with an ejection fraction of <40% was 16.7%, which was significantly greater than the 3.4% incidence of perioperative MI among patient with an ejection fraction ->40% (p -< 0.05). Other methods of identifying silent ischemia include the use of exercise tests that use radionuclide angiography or thallium scanning and the use of dipyridamole thallium testing. However, about one quarter of vascular surgical patients are unable to perform a stress test because of shortness of breath or claudication 19 and dipyridamole thallium testing is not universally available at this time. If further studies of larger groups of patients substantiate these findings and can provide an even more specific definition of the degrees of preoperative or other perioperative silent ischemia that constitute significant predictors of perioperative MI, patients shown to be at high risk could be hemodynamically monitored during the intraoperative and postoperative periods and pretreated with beta blocking agents, nitrates, and calcium channel blocking agents to limit myocardial ischemia. The efficacy of such prophylactic treatment could easily be assessed by evaluating the residual degree of ischemia by means of the iterative features of the ischemia monitor used in this study, which sounds an audible tone each time ischemia begins. In addition, the patient could be transferred to an intensive care unit for further cardiac monitoring and observation on development of a significant degree of silent ischemia during the preoperative, intraoperative, or postoperative period. Finally, if significant preoperative ischemia were detected, coronary angiography might be performed to assess the anatomic degree of coronary obstruction before any proposed peripheral vascular procedure. In conclusion, the present findings show silent myocardial ischemia to be common among patients undergoing peripheral vascular surgery. Monitoring for silent ischemia during the preoperative, intraoperative, and postoperative periods may be an effective noninvasive means of accurately identifying
Volume 10 Number 6 December 1989
those patients undergoing peripheral vascular surgical procedures who are at increased risk for a perioperative MI. Future prophylactic treatment of such patients might decrease the incidence of MI and death during the perioperative period.
Silent myocardial ischemia monitoring 625
9.
10. REFERENCES
1. Hertzer NR. Fatal myocardial infarction following abdominal aortic aneurysm resection: three hundred forty-three patients followed 6-11 years post-operatively. Ann Surg 1980;192: 667-73. 2. Hertzer NR. Fatal myocardial infarction following lower extremiry revascularization: two hundred seventy-three patients followed 6 to 11 post-operative years. Ann Surg 1981; 193:492-8. 3. Thompson JE, Austin DJ, Patman RD. Carotid endarterectomy for cerebrovascular insufficiency: long-term results in 592 patients followed up to 13 years. Ann Surg 1970; 172:663-79. 4. Hertzer NR, Lees CD. Fatal myocardial infarction following carotid endarterectomy: three hundred thirty five patients followed 6-11 years after operation. Ann Surg 1980;194: 212-8. 5. Ennix CL Jr, Lawrie GM~ Morris GC Jr, et al. Improved results of carotid endarterectomy in patients with symptomatic coronary disease: an analysis of 1546 consecutive carotid operations. Stroke 1979;10:222-5. 6. Pasternack PF, Imparato AM, Bear G, et al. The value of radionuclide angiography as a predictor ofperioperative myocardial infarction in patients undergoing abdominal aortic aneurysm resection. J VAsc SUING1984;1:320-5. 7. Pasternack PF, Imparato AM, Riles TS, et al. The value of the radionuclide angiogram in the prediction ofperioperative myocardial infarction in patients undergoing lower extremity revascularization procedures. Circulation 1985; 72 (Suppl II): II-13-17. 8. Barry J, Campbell S, Nabel EG, Mead K, Selwyn AP. Ambulatory monitoring of the digitized electrocardiogram for
11.
12.
13.
14.
15.
16.
17.
18.
19.
detection and early warning of transient myocardial ischemia in angina pectoris. Am J Cardiol 1987;60:483-8. Jamal SM, Mitra-Duncan L, Kelly DT, Freedman SB. Validation of a real-time electrocardiographic monitor for detection of myocardial ischemia secondary to coronary artery disease. Am J Cardiol 1987;60:525-7. Brown OW, Hollier LH, Pairolero PC, Kazmier FJ, McCready RA. Abdominal aortic aneurysm and coronary artery disease. Arch Surg 1981;116:1484-8. Crawford ES, Salwa AS, Babb JW, Glaeser DH, Vaccaro PS, Silver A. Infrarenal abdominal aortic aneurysm. Ann Surg 1981; 193:699- 709. Jamieson WR, Janusz MT, Miyagishima RT, Gerein AN. Influence of ischemic heart disease on early and late mortality after surgery for peripheral occlusive vascular disease. Circulation 1982;66:92-7. DeBakey ME, Crawford ES, Cooley DA, Morris GC Jr, Garrett HE, Fields WS. Cerebral arterial insufficiency: one to 11 year results following arterial reconstructive operation. Ann Surg 1965;161:921-45. Crawford ES, Bomberger RA~ Glaeser DH, Saleh SA, Russell WL. Aortoiliac occlusive disease: factors influencing survival and fimction following reconstructive operation over a twenty-five year period. Surgery 1981;90:1055-67. Burnham SJ, Johnson G, Gurri JA. Mortality risks for survivors of vascular reconstructive procedures. Surgery 1982; 92:1072-6. Hertzer NR, Beven EG, Young JR, et al. Coronary artery disease in peripheral vascular patients: a classification of 1000 coronary angiograms and results of surgical management. Ann Surg 1984;199:223-33. Holter NJ. New method for heart studies: continuous electrocardiography of active subjects over long periods is now practical. Science 1961;134:1214-20. Deanfield JE' Maseri A' Selwyn AP' et al" My°cardialischemia during daily life in patients with stable angina: its relation to symptoms and heart rate changes. Lancet 1983;1:753-8. Bhayana JN, Balu V, Hook N. Assessment of cardiac risk in surgical patients. Arch Surg 1977; 112:1448-92.
New York, N.Y. Real-time electrocardiographic monitoring for silent myocardial ischemia was performed on 200 patients undergoing peripheral vascular surgery to try to better define those at high risk of perioperative myocardial infarction. The patients were divided into those undergoing abdominal aortic aneurysm or lower extremity revascularization procedures (group I, n = 120) and those undergoing carotid artery endarterectomy (group II, n = 80). Silent ischemia was detected during the preoperative, intraoperative, or postoperative periods in 60.8% of group I and 67.5% of group II patients. Six group I and three group II patients suffered an acute perioperative myocardial infarction with two cardiac deaths. In both groups I and II a variety of parameters based on monitoring of silent myocardial ischemia were compared between the subgroups of patients who had myocardial infarction and those who did not.The results show that in both groups there was a significantly (p -< 0.05) greater total duration of perioperative ischemic time, total number of perioperative ischemic episodes, and total duration of perioperative ischemic time as a percent of total monitoring time in patients who suffered a perioperative myocardial infarction compared to those who did not. Multivariate logistic regression analysis of preoperative characteristics in all 200 patients showed the occurrence of preoperative silent myocardial ischemia and angina at rest to be the only significant predictors of perioperative myocardial infarction. Thus perioperative monitoring for silent myocardial ischemia might n0ninvasively identify those patients undergoing peripheral vascular surgery who are at increased risk for perioperative myocardial infarction, permitting implementation of timely preventive measures in selected patients. (J VASCSURG 1989;10:61725.)
Coronary artery disease is the most common cause of early postoperative death after abdominal aortic aneurysm resection, 1 lower extremity revascularization, 2 and carotid endarterectomy. 3-s Although coronary angiography can define the extent of coronary artery disease, this diagnostic approach cannot noninvasively predict which patients are most likely to suffer acute myocardial infarction (MI) and cardiac death in association with peripheral Vascular From the departmentsof Medicineand Surgery,New YorkUniversity MedicalCenter. Presented at the ThirteenthAnnualMeeting of the SouthernAssociation for Vascular Surgery, Key West, Fla., Jan 25-28, 1989. Reprint requests: Peter F. Pasternack, MD, Suite 4C, Faculty Practice, New YorkUniversityMedicalCenter, 530 First Ave., New York, NY 10016. 24/6/15572
surgical procedures. Other noninvasive methods, such as resting gated blood pool studies, 6'7 have been tested and have proved of some value in indicating those patients undergoing peripheral vascular surgery who are at the highest risk of a perioperative myocardial infarction, but no previous method has shown optimal accuracy. However, recent technologic advances now permit detection and quantitation of physiologic manifestations of myocardial ischemia in the absence of angina pectoris, a condition termed silent myocardial ischemia. Measurement of such silent myocardial ischemia might prove to be a useful noninvasive predictor of perioperative myocardial infarction in patients undergoing peripheral vascular surgical procedures. Therefore in the present study we tested the ability of a new, highly sensitive device for detecting and 617
618
Journal of VASCULAR SURGERY
Pasternack et al.
quantitating silent myocardial ischemia to distinguish the subgroup of patients at high risk for perioperative myocardial infarction in association with various peripheral vascular surgical procedures.
METHODS Patient population Two hundred patients met the criteria for admission to the study out of a total of 227 consecutive patients who underwent a peripheral vascular surgical procedure at New York University Medical Center from September 1987 to June 1988. Nineteen patients were excluded because of a QRS duration greater than 120 msec and eight patients were excluded because of left ventricular hypertrophy with marked repolarization changes. The study group included 157 men and 43 women with a mean ± SD age of 69.5 _ 8.37 years (range, 48 to 89 years). These patients had the following incidences of clinical evidence of coronary artery disease: history of exertional angina in 32%, history of rest angina in 9%, history of prior MI in 26%, and electrocardiographic (ECG) evidence of a prior MI in 14%. The 200 patients included in the study underwent one of three types of peripheral vascular surgical procedures. Sixty-seven patients (33.5%) underwent abdominal aortic aneurysm repair, 53 patients (26.5%) had lower extremity revascularization procedures, and 80 patients (40.0%) underwent carotid endarterectomy. The degree of operative stress, an important factor in the incidence of myocardial injury, differs considerably among these types of operative procedures, especially inasmuch as 75% (60/80) of the carotid endarterectomy procedures were performed with the patient under local anesthesia. Therefore the patients were subdivided into two main groups for the purposes of analysis. Group I comprised all patients undergoing abdominal aortic aneurysm or lower extremity revascularization procedures (n = 120, 60.0%). Group II contained all patients who underwent carotid endarterectomy procedures (n = 80, 40.0%). The 120 patients in group I had a mean + SD age of 70.0 _+ 8.54 years (range, 48 to 89 years) and 103 (85.8%) were men. These patients exhibited the following incidences of clinical evidence of coronary artery disease: history of exertional angina in 31.7%, history of rest angina in 7.5%, history of prior MI in 26.7%, and ECG evidence of a prior MI in 16.7%. The 80 patients in group II had a mean ± SD age of 68.9 + 8.11 years (range, 48 to 83 years) and 54 (67.5%) were men. This group manifested the following incidences of clinical evidence of coronary
artery disease: history of exertional angina in 32.5%, history of rest angina in 11.2%, history of prior MI in 25.0%, and ECG evidence of a prior MI in 11.2%.
Digital electrocardiographic monitor Patients were monitored for silent myocardial ischemia by means of a real-time ambulatory device, the QMED Monitor One Star (QMed Inc., Clark, N.J.), which continuously records and analyzes the electrocardiogram.8,9This monitor acquires the ECG signal through rigid, homogeneous, nonanodized, silver-silver chloride electrodes designed for longterm monitoring. The incoming signal is digitized at 256 Hz and passed to a 65CO2 microprocessor for analysis. Signal amplitude is compared with a fixed internal reference used for calibration. Frequency response is uniform from 0.05 to 40 Hz. The signal is analyzed by an algorithm, which is segmented into a low-level beat detection segment and a high-level analysis segment. The algorithm detects and validates a QRS complex as a region of sustained slope that meets criteria for duration, peak amplitude, and number of inflection points. Validated complexes are then analyzed to determine the onset, peak J-point, and J2 point (60 msec after J-point). The algorithm adjusts continuously to changing width of the QRS complex, and in the default condition it will not measure ST deviation in the presence of a QRS 120 msec or longer. Rapid slope changes (more than 0.15 mV in 60 msec) during the ST segment (J1 to J2) cause the device to exclude the beat from analysis for ST deviation. An instantaneous ST segment amplitude is measured at the J2 point by comparison to the amplitude of the preceding PR interval and is corrected for baseline variation. The instantaneous amplitude is used to update the ST level average on a beat-bybeat basis, and the average is stored in a series of shift registers. The definition of the amplitude of an ischemic event is selectable at - 0.1, - 0.2, - 0.3, or - 0 . 4 mV. The moment when the ST segment descends to - 0 . 1 mV is marked as the potential beginning of an ischemic event. If depression is sustained for approximately 40 seconds, the algorithm validates the ischemic event as real and begins measuring its duration. Thus time is used as a filter for accurately defining significant ST deviation. The duration of an individual depression from the point of validation to a point approximately 40 seconds after the return of the ST segment above - 0 . 1 mV is noted. The ischemic event is also segmented into durations of ST depression at - 0.1, - 0.2, - 0.3,
Volume 10 Number 6 December 1989
- 0.4 mV, and the durations of all events and segments at these levels are summed and averaged for the monitoring period. Quantitative and waveform data are stored in solid-state, random-access memory for subsequent retrieval via a liquid crystal display or a printer. A real-time clock permits the printing of time of day with each cardiac event strip. Perioperafive monitoring During the perioperative period, QMED electrodes were attached in a lead CMs configuration, with the negative reference over the sternum and positive electrode in the Vs position. The earth electrode was attached over the right lower anterior ribs. The electrodes were then connected to the recorder, and monitoring was started during the preoperative period (mean _+ SD preoperative monitoring interval: group I, 18.3 _+ 17.0 hours; group II, 1 8 . 7 _ 16.9 hours; p > 0.05) and continued throughout the operation (mean -+ SD intraoperative monitoring interval: group I, 5.46 _+ 1.92 hours, group II 3.57 -+ 1.15 hours;p < 0.001) and postoperative periods (mean --- SD postoperative monitoring interval: group I, 34.1 _+ 19.6 hours; group II, 32.4 _ 13.3 hours; p > 0.05). An ischemic ST segment change was defined as 1 mm or more of planar or downsloping ST segment depression in lead CMs. The output of the monitor consisted of indications of the presence or absence ofischemic changes and the duration and magnitude of any ST segment depression. The monitor's algorithm excludes ectopic beats and noise and identifies isoelectric, J, and J + 60 msec points in the beats classified as normal. Downsloping or horizontal ST depression of at least 1 mm leasting more than 40 seconds represents the threshold for detection of an ischemic event, but timing of event duration does not commence until the end of the 40 second validation period. The event is considered to be over when the ST amplitude has been above this 1 mm threshold for longer than 40 seconds. For each patient, up to 91 short strips of ECG representative of control and ischemic episodes were available for visual confirmation as part of the printed output. The diagnosis of acute perioperative MI was made on the basis of abnormal elevation of the serum creatine phosphokinase (CPK) MB isoenzyme level as determined daily for the first 5 postoperative days. Data collection and statistical analysis The study provided for the collection of 91 data points on each patient related to general clinical in-
Silent myocardial ischemia monitoring 619
formation such as medical history and results of diagnostic tests or to overall monitoring information. These data were obtained by reviewing the patient's hospital record or by direct questioning of the patient and by examining the record from the special ECG monitor. In addition, for each episode of silent myocardial ischemia for each patient 14 additional monitoring data points related to the ischemia, such as the time of day when the ischemia began and the duration of the episode, were gathered from the printout from the special ECG monitor. All these data were entered into a computer and analyzed by means of the statistical software packages SPSSx (SPSS Inc., Chicago, Ill.) or SAS (SAS Institute Inc., Cary, N.C.). All continuous data are expressed as mean + SD. For continuous variables the significance of differences between groups I and II was determined by the Student's t test. Within group I or II the significance of differences in continuous variables between patients who suffered an MI and patients who avoided MI was determined by a nonparametric method, the Mann-Whitney U test (Wilcoxon rank sum W test). For categoric variables the significance of any differences between the group that suffered a perioperative MI and the group that avoided an MI was analyzed by the chi-square test. Logistic regression analysis was used for multivariate analysis of categoric variables to determine the best set of variables independently predictive of perioperative MI. The estimate of the coefficient in the regression equation, the estimate of the odds ratio, and the 95% confidence interval for the odds ratio were maximum likelihood estimates. The odds ratio is the relative odds of the occurrence of a perioperative MI associated with a change in the predictor variable. RESULTS Silent ischemia Overall, the occurrence of one or more episodes of silent myocardial ischemia was detected in 127 patients (63.5%) during the preoperative, intraoperative, or postoperative periods. In group I, one or more episodes of silent myocardial ischemia appeared in 73 patients (61%) during one or more of these periods. During a period of preoperative monitoring of 21.2 + 19.3 hours, 48 patients (40%) showed evidence of silent ischemia. Among these patients the number of preoperative ischemic episodes was 13.0 + 16.7, the total duration of preoperative ischemia was 5.4 + 8.2 hours, and the total duration of preoperative silent ischemia as a per-
Journal of VASCULAR
620 Pasternack et aI.
SURGERY
Table I. Characteristics of silent myocardial ischemia among 120 patients with abdominal aortic aneurysm or lower extremity revascularization (group I) who suffered myocardial ischemia~ Preoperative Intraoperative Postoperative No. of patients with ischemia (%) Duration of monitorhag (hr) Duration ofischemia (hr) No. ofischemic episodes Total duration of silent myocardial ischemia as % of monitoring interval
48 (40.0%)
45 (37,5%)
57 (47.5%)
21.2 -+ 19.3
5.40 _+ 1.60
38.6 + 19.2
5.37 + 8.18
1.58 + 2.38
7.13 -+ 11.6
13.0 + 16.7
4.44 _+_4.10
12.7 + 13.7
29.0 + 36.2%
32.4 _+ 49.6%
20.5 _+ 30.9%
~Patients had myocardial ischemia during the preoperative, intraoperative, or postoperative periods (mean _+ SD).
Table II. Characteristics of silent myocardial ischemia among 80 patients with carotid artery endarterectomy (group II) who suffered myocardial ischemia~ No. of patients with ischemia (%) Duration of monitoring (hr) Duration ofischemia (hr) No. ofischemic episodes Total duration of silent myocardial ischemia as % of monitoring interval
Preoperative
Intraoperative
Postoperative
30 (37.5%)
33 (41.2%)
43 (53.8%)
22.9 _+ 21.1
3.82 -+ 1.61
32.9 + 11.8
5.16 _+ 11.0
0.65 + 0.85
4.74 +_ 8.47
15.4 _+ 19.3
3.36 _+ 2.74
17,6 _+ 19.4
21.9 +_ 30.8%
19.7 _+ 26.4%
14.3 + 24.8%
*Patients had myocardial ischemia during preoperative, intraoperative, or postoperative periods (mean _+ SD).
centage of the total duration of preoperative monitoring time (percent time ischemic) was 29.0% _+ 36.2% (Table I). During a period of intraoperative monitoring of 5.4 _+ 1.6 hours, 45 patients (38%) showed evidence of silent ischemia. In these patients the number of intraoperative ischemia episodes was 4.4 _+ 4.1, the total duration of intraoperative ischemia was 1.6 _+ 2.4 hours, and the percent time ischemic was 32.4% _+ 49.6%. During a period of postoperative monitoring of 38.6 _+ 19.2 hours, 57 patients (48%) showed evidence of silent ischemia. Among these patients the number of postoperative ischemia episodes was 12.7 _+ 13.7, the total duration of postoperative ischemia was 7.1 _+ 11.6 hours, and the percent time ischemic was 20.5% _+ 30.9%. In group II, one or more episodes of silent myocardial ischemia were found in 54 patients (68%)
during the preoperative, intraoperative, or postoperative periods. During a period of preoperative monitoring of 22.9 _+ 21.1 hours, 30 patients (38%) had evidence of silent ischemia. Among these patients the number of preoperative ischemic episodes was 15.4 _+ 19.3, the total duration of preoperative ischemia was 5.2 + 11.0 hours, and the percent time ischemic was 21.9% + 30.8% (Table II). During a period of intraoperative monitoring of 3.8 _+ 1.6 hours, 33 patients (41%) showed evidence of silent ischemia. In these patients the number of intraoperative ischemic episodes was 3.4 _+ 2.7, the total duration of intraoperative ischemia was 0.65 _ 0.85 hours, and the percent time ischemia was 1 9 . 7 % +_ 26.4%. During a period of postoperative monitoring of 32.9 _ 11.8 hours, 43 patients (54%) showed evidence of silent ischemia. Among these patients the number of postoperative ischemic episodes was
Volume 10 Number 6 December 1989
Silent myocardial ischemia monitoring
621
Table III. Indexes of silent ischemia among 120 patients with abdominal aortic aneurysms or lower extremity revascularization (group I) + Myocardial infarction No myocardialinfarction p value No. o f patients Time monitored preoperatively (hr) Time monitored intraoperatively (hr) Time monitored postoperatively (hr) Total time monitored perioperatively (hr) Duration o f preoperative SI (hr) Duration o f intraoperative SI (hr) Duration ofpostoperatire SI (hr) Total duration ofperioperative SI (hr) No. o f preoperative SI episodes No. ofintraoperative SI episodes N o o f postoperative SI episodes Total no. ofperioperative SI episodes Duration o f SI as % o f preoperative monitoring interval Duration o f SI as % o f intraoperative monitoring interval Duration o f SI as % o f postoperative monitoring interval Total duration o f SI as % o f total perioperative monitoring interval
6 21.8 +- 21.0
114 18.2 -- 16.9
-NS
5.78 -4- 2.36
5.44 _+ 1.90
NS
46.8 z 19,6
33.4 -+ 19.4
NS
74.4 -4- 34.0
56.7 _+ 25.2
NS
3.73 _+ 4.07
2.06 -+ 5.86
0.036
2,40 + 4.05
0.50 -- 1.39
NS
11.5 + 13.1
2.96 _+ 8.32
0.001
I7.7-+ 17.8
5.52 + 13.5
0.002
11.5 + 19.0
5.49 +_ 11.6
0.014
4.50 -~ 5.05
2.18 + 2.84
NS
26.2 + 17.2
5.54 +_ 9.74
0.001
42.2 + 25,4
13,2 + 18.1
0,001
31.2 -+ 41.6%
10.6 +_ 25.7%
0,035
30.4 -+ 44.1%
11.2 + 33.4%
NS
23.8 -+ 25.4%
9.02 + 23.3%
0.001
24.0 +_ 26.3%
9.84 -+ 23.8%
0.003
NS, Not significant; SI, silent myocardial ischemia; mean + SD. +Includes patients with myocardial infarction and those who avoided myocardial infarction.
17.6 + 19.4, the total duration of postoperative ischemia was 4.7 ___ 8.5 hours, and the percent time ischemic was 14.3% _+ 24.8%. Myocardial infarction Six patients in group I (three patients with abdominal aortic aneurysm and three patients with' lower extremity revascularization) and three in group II suffered acute perioperative MIs with CPK-MB levels present in the serum within the first 48 hours after surgery. All nine patients displayed evidence of silent ischemia during the preoperative, intraoperative, or postoperative periods, and there were no perioperative MIs among patients who escaped perioperative silent myocardial ischemia. There were two cardiac deaths among the 200 patients, both in patients in group I who had shown
evidence of silent myocardial ischemia. One suffered an extension of non-Q wave MI and died on the twelfth postoperative day. The other patient, who did not suffer a perioperative MI but who had evidence of preoperative, intraoperative, and postoperative ischemia, had refractory congestive heart failure and died in ventricular fibrillation on the twelfth postoperative day. In group I there was no significant difference in preoperative, intraoperative, postoperative, or total perioperative time monitored between those patients who suffered a perioperative MI and those with no MI (Table III). However, the total duration of pcrioperative silent ischemia, the total number of perioperative episodes of silent ischemia, and the total duration of perioperativc silent ischemia as a percentage of total perioperative time monitored were
622
Journal of VASCULAR SURGERY
P a s t e r n a c k e t al.
T a b l e I V . Indexes o f silent ischemia a m o n g 80 patients with carotid artery endarterectomy (group II) *
No. of patients Time monitored preoperatively (hr) Time monitored intraoperatively (hr) Time monitored postoperatively (hr) Total time monitored perioperatively Total duration ofpreoperative SI (hr) Total duration ofintraoperative SI (hr) Total duration ofpostoperative SI (hr) Total duration of perioperative SI (hr) No. of preoperative SI episodes No. ofintraoperative SI episodes No. of postoperative SI episodes Total no. ofpetioperative episodes Total duration of SI as % of preoperative monitoring interval Total duration of SI as % of intraoperative monitoring interval Total duration of SI as % of postoperative monitoring interval Total duration of SI as % of total perioperative monitoring interval
Myocardial infarction
No myocardial infarction
p value
3 14.0 -+ 1.83
77
--
18.8 +_ 17.2
NS
3.94 -- 0.82
3.55 + 1.16
NS
38.3 _+ 13.1
32.2 -+ 13.3
NS
53.1 --- 18.9
56.2 + 14.6
NS
0.04 + 0.03
2.01 - 7.26
NS
0.42 _+ 0.20
0.26 -+ 0.64
0.033
6.78 _+ 5.68
2.38 +- 6.63
0.016
7.24 + 5.60
4.65 -+ 12.7
0.050
2.00 _+ 1.00
6.57 -+ 13.9
NS
3.33 + 1.53
1.92 + 2.11
0.011
39.3 + 23.5
8.79 + 15.2
0.009
44.7 _+ 23.5
17.3 + 24.0
0.029
0.24 _+ 0.23%
8.51 -+ 21.8%
NS
10.6 _+ 3.89%
8.04 -+ 19.8%
0.037
16.1 _+ 8.15%
7.39 + 19.7%
0.023
11.9 _+ 6.07%
7.43 -+ 18.2%
0.050
NS, Not significant; SI, silent myocardial ischemia; mean _+ SD. *Includes patients with myocardial infarction and those who avoided myocardial infarction.
significantly greater in those patients w h o suffered an acute perioperative M I c o m p a r e d to those with n o MI. T h e patients w h o suffered an acute M I s h o w e d a total duration o f perioperative silent myocardial ischemia o f 17.7 -- 17.8 hours c o m p a r e d to 5.52 _ 13.5 hours for those patients with n o M I (p = 0.002). Likewise, the patients w h o suffered an M I experienced a total o f 42.2 _+ 25.4 episodes
of perioperative silent ischemia compared to a total o f 13.2-_+ 18.1 episodes in those with n o M I (p = 0.001). In addition, the total duration o f p e r i operative silent ischemia as a percentage o f the total perioperative time m o n i t o r e d (percent time ischemic) was significantly higher in those suffering an acute perioperative M I c o m p a r e d to those with no MI. T h e patients w h o demonstrated elevation o f C P K - M B had a perioperative percent time ischemic
o f 24.0% --_ 26.3% c o m p a r e d to 9.84% + 23.8% for those with n o M I (p = 0.003). Similarly, in g r o u p II there was n o significant difference in preoperative, intraoperative, postoperative, or total perioperative time m o n i t o r e d between those patients w h o suffered a perioperative M I and those with n o M I (Table IV). H o w e v e r , the total duration o f perioperative silent ischemia, the total n u m b e r o f perioperative episodes o f silent ischemia, and the total duration ofperioperative silent ischemia as a percentage o f total perioperative time m o n i t o r e d were significantly greater in those patients w h o suffered an acute perioperative M I c o m p a r e d to those with no MI. The patients w h o suffered an acute M I s h o w e d a total duration o f perioperative silent myocardial ischemia o f 7.24 _+ 5.60 hours c o m p a r e d to 4.65 _+ 12.7 hours for those patients with n o M I
Volume 10 Number 6 December 1989
(p < 0.05). Also, the patients who suffered an MI experienced a total of 44.7 _+ 23.5 episodes of perioperative silent ischemia compared to a total of 17.3 + 24.0 episodes in those with no MI (p = 0.03). Finally, the patients who demonstrated elevation of CPK-MB had a perioperative percent time ischemic of 11.9% + 6.07% compared to 7.43% __+ 18.2% for those with no MI (p -< 0.05). Multivariate analysis of categoric perioperative factors was performed with a logistic regression model to determine which factors might be significant predictors of perioperative MI. In addition to the occurrence of preoperative silent myocardial ischemia, the variables tested included a history of a previous MI, hypertension, angina at rest, angina on exertion, congestive heart failure, and EGG evidence of a previous MI. When the patients were divided into the relatively small groups I and II and these were tested separately, group I showed only the presence of angina at rest to be a significant predictor of perioperative MI (beta coefficient in logit model = 2.89, p = 0.002), with an odds ratio of 18.0 (95% confidence interval = 2.97, 109). In group II none of the variables tested proved to be a significant predictor of MI. However, testing of the entire group of 200 patients undergoing any of three types of peripheral vascular surgery showed the presence of preoperative silent myocardial ischemia (beta = 1.56, p = 0.04) and angina at rest (beta = 1.66, p = 0.04) to be significant predictors of perioperative MI with odds ratios of 4.76 (95% confidence interval = 1.03, 22.0) and 5.2 (95% confidence interval = 1.04, 22.6), respectively. DISCUSSION As noted, coronary artery disease remains the leading cause of early and late mortality after peripheral vascular surgical procedures. Acute myocardial infarction is the cause of about one half of all postoperative deaths after abdominal aortic aneurysm repair, 1°12 carotid endarterectomy,saa or lower extremity revasculatization procedures. 12'14 Longterm survival is also substantially limited by a high incidence of late cardiac deaths, especially among those with obvious evidence of coronary artery disease at the time of their initial presentationY -~s Thus an easily accessible, inexpensive, noninvasive method of identifying those patients who are likely to suffer or even die of an MI in association with peripheral vascular surgical procedures would be extremely helpful in perioperative management. The digitized real-time ischemia monitor used in this
Silent myocardial ischemia monitoring 623
study appears to be a promising candidate to fill this role. The rcsuks of this study show that there is a previously unappreciated, extensive amount of silent myocardial ischemia present in patients undergoing peripheral vascular surgical procedures, with 63.5% of 200 patients (60.8% in group I, 67.5% in group II) suffering silent myocardial ischemia during the preoperative, intraoperative, or postoperative period. It is likely that this high incidence of silent ischemia has an anatomic correlation and is a manifestation of severe underlying coronary artery disease in this patient population. In an earlier study cardiac catheterization of 1000 patients undergoing abdominal aortic aneurysm surgery, lower extremity revascularization, or carotid endartcrectomy demonstrated significant coronary artery disease (>70% stenosis) involving a single coronary artery in 27%, two coronary arteries in 19%, and three coronary arteries in 11% of these patients. 16 Recent technologic advances have rapidly accelerated the accumulation of knowledge regarding silent myocardial ischemia. In 1961 Holter 17observed and described an apparent episode of silent ischemia in a patient known to have angina. Changes observed on an ambulatory ECG recording were equivalent to those sccn. during angina but were not associated with pain. Since then the development of accurate monitoring of the ST segment has shown that patients with coronary artery disease have frequent episodes of such painless myocardial ischemia during daily activity. In fact, 75% of all ischemia episodes have been found to be painless. Is Such silent ischemia can occur even at low heart rates and in the absence of physical exercise. Thus silent ischemia is simply myocardial ischemia that occurs in the absence of chest pain and results from the same sequence of physiologic events as ischemia accompanied by angina. Conjecture as to why myocardial ischemia is so often silent has involved several hypotheses including: silent ischemia indicates less severe ischemia of a degree insufficient to reach pain thresholds; there are differences in pain perception thresholds or central transmission of painful stimuli among individuals; and there may be different pathophysiologic characteristics resulting in different patterns of myocardial blood flow distribution during silent, as opposed to symptomatic, ischemia. In addition, there has been much recent speculation regarding the potential prognostic implications, if any, of the common phenomenon of silent
624
Pasternack et al.
ischemia. The data obtained from this study suggest that the frequency and duration of perioperative episodes of silent ischemia are significantly associated with a higher incidence of acute perioperative MI. Three different primary indexes of the severity of silent ischemia were tested for groups I and II in this study: the total duration of perioperative ischemia, the total number of perioperative ischemic episodes, and the total perioperative percent time ischemic. In both groups the total duration of episodes of perioperative silent ischemia was found to be significantly longer in patients suffering an acute perioperative MI compared to those with no MI (Tables III & IV). In addition, in both groups I and II the total number of perioperative ischemic episodes and the total perioperative percent time ischemic were significantly longer in those suffering an acute perioperative MI compared to those with no MI (Tables III & IV). Thus the present findings show that perioperative silent ischemia is of significantly longer total duration and significantly more frequent in those patients who suffer acute perioperative MI. The longer total duration and more numerous episodes of perioperative ischemia in patients who suffer MI may play a role in creating irreversible myocardial damage during the perioperative period. For the total study group of 200 patients multivariate analysis of most available factors that might be considered predictors of perioperative MI, including a history of previous MI, hypertension, angina at rest, angina on exertion, congestive heart failure, ECG evidence of previous MI, and the presence of preoperative silent myocardial ischemia, showed only angina at rest and preoperative silent myocardial ischemia to be significant predictors of perioperative MI. A patient with angina at rest had 4.7 times the risk of a perioperative MI in association with peripheral vascular surgery as a patient without angina at rest. A patient who manifested silent myocardial ischemia preoperatively had 5.2 times the risk of a perioperative MI as a patient without preoperative silent myocardial ischemia. However, it should be noted that since the model used for logistic regression analysis is exploratory in nature, any results regarding the predictive value of certain variables with respect to the occurrence of perioperative MI should be validated by further studies. These findings regarding the predictive value of preoperative silent ischemia in patients who are to undergo peripheral vascular surgery is encouraging with respect to the goal of timely identification of those patients who are more likely to suffer an acute perioperative MI. In this study preoperative moni-
Journal of VASCULAR SURGERY
toring was initiated less than 24 hours before surgery, a lead time that should be of practical value for early identification of patients with significant evidence of preoperative silent ischemia. As we have noted in previous work, 6,7 those patients undergoing peripheral vascular surgery who have low cardiac ejection fractions as determined by preoperative resting gated blood pool studies have a significantly higher incidence of perioperative MI than patients with higher ejection fractions. Similarly analyzed data from the present study of a total of 200 patients showed that the incidence of perioperative MI among patients with an ejection fraction of <40% was 16.7%, which was significantly greater than the 3.4% incidence of perioperative MI among patient with an ejection fraction ->40% (p -< 0.05). Other methods of identifying silent ischemia include the use of exercise tests that use radionuclide angiography or thallium scanning and the use of dipyridamole thallium testing. However, about one quarter of vascular surgical patients are unable to perform a stress test because of shortness of breath or claudication 19 and dipyridamole thallium testing is not universally available at this time. If further studies of larger groups of patients substantiate these findings and can provide an even more specific definition of the degrees of preoperative or other perioperative silent ischemia that constitute significant predictors of perioperative MI, patients shown to be at high risk could be hemodynamically monitored during the intraoperative and postoperative periods and pretreated with beta blocking agents, nitrates, and calcium channel blocking agents to limit myocardial ischemia. The efficacy of such prophylactic treatment could easily be assessed by evaluating the residual degree of ischemia by means of the iterative features of the ischemia monitor used in this study, which sounds an audible tone each time ischemia begins. In addition, the patient could be transferred to an intensive care unit for further cardiac monitoring and observation on development of a significant degree of silent ischemia during the preoperative, intraoperative, or postoperative period. Finally, if significant preoperative ischemia were detected, coronary angiography might be performed to assess the anatomic degree of coronary obstruction before any proposed peripheral vascular procedure. In conclusion, the present findings show silent myocardial ischemia to be common among patients undergoing peripheral vascular surgery. Monitoring for silent ischemia during the preoperative, intraoperative, and postoperative periods may be an effective noninvasive means of accurately identifying
Volume 10 Number 6 December 1989
those patients undergoing peripheral vascular surgical procedures who are at increased risk for a perioperative MI. Future prophylactic treatment of such patients might decrease the incidence of MI and death during the perioperative period.
Silent myocardial ischemia monitoring 625
9.
10. REFERENCES
1. Hertzer NR. Fatal myocardial infarction following abdominal aortic aneurysm resection: three hundred forty-three patients followed 6-11 years post-operatively. Ann Surg 1980;192: 667-73. 2. Hertzer NR. Fatal myocardial infarction following lower extremiry revascularization: two hundred seventy-three patients followed 6 to 11 post-operative years. Ann Surg 1981; 193:492-8. 3. Thompson JE, Austin DJ, Patman RD. Carotid endarterectomy for cerebrovascular insufficiency: long-term results in 592 patients followed up to 13 years. Ann Surg 1970; 172:663-79. 4. Hertzer NR, Lees CD. Fatal myocardial infarction following carotid endarterectomy: three hundred thirty five patients followed 6-11 years after operation. Ann Surg 1980;194: 212-8. 5. Ennix CL Jr, Lawrie GM~ Morris GC Jr, et al. Improved results of carotid endarterectomy in patients with symptomatic coronary disease: an analysis of 1546 consecutive carotid operations. Stroke 1979;10:222-5. 6. Pasternack PF, Imparato AM, Bear G, et al. The value of radionuclide angiography as a predictor ofperioperative myocardial infarction in patients undergoing abdominal aortic aneurysm resection. J VAsc SUING1984;1:320-5. 7. Pasternack PF, Imparato AM, Riles TS, et al. The value of the radionuclide angiogram in the prediction ofperioperative myocardial infarction in patients undergoing lower extremity revascularization procedures. Circulation 1985; 72 (Suppl II): II-13-17. 8. Barry J, Campbell S, Nabel EG, Mead K, Selwyn AP. Ambulatory monitoring of the digitized electrocardiogram for
11.
12.
13.
14.
15.
16.
17.
18.
19.
detection and early warning of transient myocardial ischemia in angina pectoris. Am J Cardiol 1987;60:483-8. Jamal SM, Mitra-Duncan L, Kelly DT, Freedman SB. Validation of a real-time electrocardiographic monitor for detection of myocardial ischemia secondary to coronary artery disease. Am J Cardiol 1987;60:525-7. Brown OW, Hollier LH, Pairolero PC, Kazmier FJ, McCready RA. Abdominal aortic aneurysm and coronary artery disease. Arch Surg 1981;116:1484-8. Crawford ES, Salwa AS, Babb JW, Glaeser DH, Vaccaro PS, Silver A. Infrarenal abdominal aortic aneurysm. Ann Surg 1981; 193:699- 709. Jamieson WR, Janusz MT, Miyagishima RT, Gerein AN. Influence of ischemic heart disease on early and late mortality after surgery for peripheral occlusive vascular disease. Circulation 1982;66:92-7. DeBakey ME, Crawford ES, Cooley DA, Morris GC Jr, Garrett HE, Fields WS. Cerebral arterial insufficiency: one to 11 year results following arterial reconstructive operation. Ann Surg 1965;161:921-45. Crawford ES, Bomberger RA~ Glaeser DH, Saleh SA, Russell WL. Aortoiliac occlusive disease: factors influencing survival and fimction following reconstructive operation over a twenty-five year period. Surgery 1981;90:1055-67. Burnham SJ, Johnson G, Gurri JA. Mortality risks for survivors of vascular reconstructive procedures. Surgery 1982; 92:1072-6. Hertzer NR, Beven EG, Young JR, et al. Coronary artery disease in peripheral vascular patients: a classification of 1000 coronary angiograms and results of surgical management. Ann Surg 1984;199:223-33. Holter NJ. New method for heart studies: continuous electrocardiography of active subjects over long periods is now practical. Science 1961;134:1214-20. Deanfield JE' Maseri A' Selwyn AP' et al" My°cardialischemia during daily life in patients with stable angina: its relation to symptoms and heart rate changes. Lancet 1983;1:753-8. Bhayana JN, Balu V, Hook N. Assessment of cardiac risk in surgical patients. Arch Surg 1977; 112:1448-92.