Author + information
- Received June 20, 2000
- Revision received January 30, 2001
- Accepted February 13, 2001
- Published online June 1, 2001.
- Giora Landesberg, MD, DSc∗,* (, )
- Morris Mosseri, MD†,
- Doron Zahger, MD§,
- Yehuda Wolf, MD‡,
- Misha Perouansky, MD∗,
- Haim Anner, MD‡,
- Benjamin Drenger, MD∗,
- Yonatan Hasin, MD†,
- Yacov Berlatzky, MD‡ and
- Charles Weissman, MD∗
- ↵*Reprint requests and correspondence: Dr. Giora Landesberg, Department of Anesthesiology and Critical Care Medicine, Hadassah University Hospital Kiryat Hadassah, Ein Kerem, Jerusalem, Israel 91120
The goal of this study was to investigate the nature of the association between silent ischemia and postoperative myocardial infarction (PMI).
Silent ischemia predicts cardiac morbidity and mortality in both ambulatory and postoperative patients. Whether silent stress-induced ischemia is merely a marker of extensive coronary artery disease or has a closer association with infarction has not been determined.
In 185 consecutive patients undergoing vascular surgery, we correlated ischemia duration, as detected on a continuous 12-lead ST-trend monitoring during the period 48 h to 72 h after surgery, with cardiac troponin-I (cTn-I) measured in the first three postoperative days and with postoperative cardiac outcome. Postoperative myocardial infarction was defined as cTn-I >3.1 ng/ml accompanied by either typical symptoms or new ischemic electrocardiogram (ECG) findings.
During 11,132 patient-hours of monitoring, 38 patients (20.5%) had 66 transient ischemic events, all but one denoted by ST-segment depression. Twelve patients (6.5%) sustained PMI; one of those patients died. All infarctions were non-Q-wave and were detected by a rise in cTn-I during or immediately after prolonged, ST depression-type ischemia. The average duration of ischemia in patients with PMI was 226 ± 164 min (range: 29 to 625), compared with 38 ± 26 min (p = 0.0000) in 26 patients with ischemia but not infarction. Peak cTn-I strongly correlated with the longest, as well as cumulative, ischemia duration (r = 0.83 and r = 0.78, respectively). Ischemic ECG changes were completely reversible in all but one patient who had persistent new T wave inversion. All ischemic events culminating in PMI were preceded by an increase in heart rate (Δ heart rate = 32 ± 15 beats/min), and most (67%) of them began at the end of surgery and emergence from anesthesia.
Prolonged, ST depression-type ischemia progresses to MI and is strongly associated with the majority of cardiac complications after vascular surgery.
Silent Holter-detected ischemia is associated with cardiac morbidity and mortality both in ambulatory patients with coronary artery disease (CAD) (1,2)and in high-risk cardiac patients undergoing noncardiac surgery (3,4). A number of investigators have attempted to improve cardiac outcome by prophylactic therapy to prevent silent ambulatory or postoperative ischemia (5–7). Nevertheless, whether silent ischemia is just a marker for more extensive CAD or may also play a mechanistic role in the evolution of myocardial infarction (MI) is debatable (8,9). The elusiveness of this issue is at least, in part, related to the relatively low sensitivity (37% to 50%) and specificity (88% to 92%) of Holter monitoring in detecting myocardial ischemia (10,11)and the relative inaccuracy of creatine kinase (CK)-MB isoenzyme in identifying postoperative MI (PMI) (12).
Patients undergoing vascular surgery have a high prevalence (50% to 60%) of CAD and are at a particularly high risk for PMI and cardiac death (5% to 15%) (13). Nevertheless, despite numerous studies that investigated almost every predictor for PMI, its mechanisms remain poorly understood (14,15). This study correlates data from continuous perioperative 12-lead ST-segment monitoring with serial cardiac troponin-I (cTn-I) measurements and postoperative cardiac outcome in an attempt to obtain a better insight into the nature of the association between postoperative ischemia and PMI after major vascular surgery.
Patients and methods
After the approval of the institutional review board and after receiving informed consent, 185 consecutive patients undergoing major vascular surgery (84: carotid endarterectomy, 28: abdominal aortic surgery and 73: lower-extremity bypass procedure) at the Hadassah University Hospital were studied. Patients with unstable angina or MI within the preceding three months were excluded. Perioperative clinical data were recorded prospectively. The preoperative 12-lead electrocardiogram (ECG) was analyzed based on the Sokolow-Lyon criteria for left ventricular hypertrophy (LVH) as previously described (16). Monitoring included continuous intra-arterial blood pressure and pulse oximetry measurement during surgery and for at least another day. Seven patients had regional (epidural or continuous spinal) anesthesia; 65 had combined general and epidural anesthesia, and 113 patients had general anesthesia only. After completion of surgery, patients were treated in the recovery room or intensive care unit at least until the morning after surgery. All preoperative medications were resumed postoperatively as soon as the patients were able to take fluids by mouth. All cardiac signs and symptoms during the hospitalization were recorded, and a 12-lead ECG was obtained before hospital discharge.
Continuous 12-lead ECG recording
Before induction of anesthesia, patients were connected to a continuous 12-lead ECG monitor (Solar 7000, Marquette Electronics, Milwaukee, Wisconsin) wired through a network to a Cardiac Review Station (ST-Guard). Each minute the ST-Guard stored all 12-lead ECG complexes, measured the ST-segment deviation in all leads compared with the baseline ECG and displayed the ST trends. The ST segment was measured 60 ms after the J point, and an episode of ST deviation was defined as ST depression or elevation of ≥0.2 mV in one lead or ≥0.1 mV in two contiguous leads that lasted >10 min. Each episode of ST-segment deviation was automatically detected and marked by the ST-Guard. Monitoring was continued for 72 h, except in carotid endarterectomy patients who usually ambulated after 48 h. Treating physicians were blinded to the ST-Guard data. However, if ischemia was clinically suspected based on other monitors or clinical signs, physicians were allowed to examine the changes on ST-Guard and treat those patients accordingly. Treatment to reverse ischemia included: improving oxygenation, intravenous beta-adrenergic blocking agents (esmolol/labetalol), nitroglycerine and correction of anemia by blood transfusion, as clinically indicated. The 12-lead ST-segment trends were reviewed by the study investigators, and artifacts were deleted. Periods marked by the ST-Guard as ST-segment deviations were inspected visually for accuracy, and those reflecting artifacts or pure up-sloping ST-segment depression were not considered as ischemia. Each patient’s longest and cumulative ischemia duration was recorded. Mean heart rate during the 30 min before, at the onset of and at maximum ST deviation was recorded from the ST-Guard.
Cardiac troponin-I and CK-MB were measured in all patients immediately after surgery daily for the first three postoperative days and later if clinically indicated. Cardiac troponin-I was measured using a Stratus II analyzer (Dade International, Deerfield, Illinois) by mass immunoassay with two monoclonal specific antibodies. Serum total CK and its MB isoenzyme activity were measured by a Kodak Ektachem multiple-point rate assay (Johnson & Johnson Co., New Brunswick, New Jersey).
Cardiac death was defined as death secondary to MI, arrhythmia or congestive heart failure. Myocardial infarction was defined as an increase in cTn-I >3.1 ng/ml (12,17)accompanied by at least one of the following: typical ischemic symptoms, ECG changes indicative of ischemia (ST-segment depression or elevation) or new pathological Q waves. Cardiac troponin-I >3.1 ng/ml is also above the 99th percentile of a reference population in our lab. This definition of MI was based on the new consensus document formulated by the Joint American-European Task Force for Redefinition of MI (18). All cardiac events were assessed and agreed upon by at least two of the investigators.
Student ttest and chi-square analyses were used to compare variables between groups of patients. Two-tailed Pearson’s correlation coefficient was used to correlate ischemia duration with cTn-I and CK-MB. Receiver-operator characteristic curve (ROC) analysis was used to assess the concordance between cTn-I and ischemia duration and to define the cutoff values of these variables associated with largest area under the ROC. All preoperative predictors of MI with p value <0.1 on univariate analysis were included in a multivariate logistic regression analysis. Forward stepwise conditional selection method (probability criteria for stepwise: entry = 0.05, removal = 0.1) identified the variables independently associated with PMI.
Table 1summarizes the preoperative clinical features of all patients and demonstrates the relatively high incidence of CAD and cardiovascular risk factor typical to these patients. During 11,132 patient-hours of monitoring (60.7 ± 11.9 per patient), 38 patients (20.5%) had 66 ischemic episodes (1.7 ± 1.4 per patient, range: 1 to 8), all but one denoted by ST-segment depression. One patient had an episode of ST-segment elevation in leads L2, L3and aVF associated with ST-segment depression in V1to V3and aVL that lasted 24 min and was not accompanied by an elevation of cardiac markers. Table 2depicts the characteristics of myocardial ischemia and infarction in the 38 patients with ischemia. The duration of the longest ischemic events was 96 ± 127 min (range: 11 to 625 min), and the cumulative ischemia duration over the entire monitoring period was 150 ± 252 min (range: 11 to 1,150 min). In seven patients, myocardial ischemia was clinically suspected by the treating physicians and their ST-Guard data were disclosed in order to verify and treat ischemia. Four of these patients already had prolonged ST-segment depression on the ST-Guard at the time ischemia was clinically suspected and subsequently had PMI.
Twelve patients (6.5%) sustained MI (cTn-I levels: 21.1 ± 26.5 ng/ml, range: 3.3 to 100.2). All infarctions were non-Q-type. In all 12 patients, the increase in cTn-I was detected either during (two patients) or within 18 h from a prolonged, transient ST-segment depression. All seven clinically evident cardiac events occurred in these patients: five patients had prolonged chest pain during peak ischemia, one of whom died from cardiac arrest seven days postoperatively; two patients developed congestive heart failure during the prolonged ischemia. In the other five patients, MI was asymptomatic. None of the 185 patients had new Q waves. In all patients with ischemia, including those with infarction, the ECG changes reverted completely to baseline, except for one patient in whom new T wave inversion persisted for >48 h. None of the 173 patients without MI suffered prolonged chest pain, cardiac failure or cardiac death during the hospitalization.
Peak serum cTn-I concentration strongly correlated with both longest ischemia and cumulative ischemia duration (r = 0.83, p < 0.0001 and r = 0.78, p < 0.0001, respectively) (Fig. 1). In contrast, maximal CK-MB correlated poorly with longest or cumulative ischemia duration (r = 0.25 and r = 0.19, respectively).
Cardiac troponin-I >3.5 ng/ml was associated with the maximal area under the longest ischemia duration ROC (Fig. 2)and close to maximal area (0.923) under the CK-MB ROC. Similarly, longest ischemia duration of >120 min was associated with maximal area under the cTn-I ROC (Fig. 2).
Patients who sustained MI had more than fivefold longer episodes of ischemia, and the cumulative ischemia duration was more than sevenfold longer, compared with patients with ischemia but not MI (Table 2). Of 13 patients with longest ischemia episodes of >100 min, only two had cTn-I levels below 3.1 ng/ml (Fig. 1), and in only one of the patients with cTn-I >3.1 ng/ml was the longest ischemia <100 min. Ten (83%) of the patients with infarction had elevated CK-MB as well.
All ischemic events culminating in PMI were preceded by an increase in heart rate. In the 38 patients with ischemia, heart rate during the 30 min before ischemia was 84 ± 12 beats/min (range: 60 to 108). It increased to 106 ± 18 beats/min (range: 71 to 160, p < 0.0001) at the onset of ischemia and to 116 ± 18 beats/min (range: 82 to 162, p < 0.0001) at the time of maximal ST-segment deviation. Among patients with ischemia, there was no difference in heart rate during ischemia between patients with and without MI (Table 2). Blood pressure at onset of ischemia in patients with MI was within their preoperative range, except for one patient who was mildly hypotensive (90/60 mm Hg). Five (45%) patients with MI had hemoglobin concentration <10 gr/dl, and, in all but one patient, the hemoglobin concentration was lower during ischemia than it was before surgery (Δ hemoglobin = 2.9 ± 1.5 gr/dl). There was no difference, however, between patients with and without MI in either the postoperative hemoglobin concentration (11.3 ± 1.5 gr/dl and 11.3 ± 1.4 gr/dl, respectively) or its decrement from the preoperative value (Δ hemoglobin = 1.9 ± 1.3 gr/dl and 1.6 ± 1.6 gr/dl, respectively, p = NS).
Preoperative predictors of MI
Table 1depicts the preoperative predictors of perioperative MI. By multivariate logistic regression analysis, diabetes mellitus and LVH were the only independent predictors of PMI (odds ratio = 4.1, 95% confidence interval [CI] = 1.2 to 14.4, p = 0.027; odds ratio = 3.77, 95% CI = 1.1 to 12.7, p = 0.032, respectively).
Timing of myocardial ischemia and infarction
Twenty-six (68%) of all the longest ischemic events started during the period between 50 min before and 60 min after the end of surgery, during emergence from anesthesia. In eight (67%) of the patients with PMI, the longest ischemic episodes started within the same time period (Fig. 3), and their cTn-I reached >3.1 ng/ml within less than 18 h. In the other four patients with PMI, the rise in cTn-I >3.1 ng/ml occurred in the second and third postoperative days.
No patient had ischemia (>10 min) during induction of anesthesia. Six patients had relatively short (≤34 min) intraoperative ischemia; only one of them sustained a longer episode of postoperative ischemia, which culminated in MI.
Previous studies implied that prolonged ischemia is associated with postoperative cardiac complications. Frank et al. (19)and Ganz et al. (20)described two patients who died 9 h and 12 h after peripheral vascular surgery, following prolonged, silent ST depression-type ischemia on Holter monitoring. In a study of 151 patients, the cumulative duration of ST depression on Holter monitoring predicted postoperative cardiac complications (4). This study expands previous observations and shows that in vascular surgery patients the rise in cTn-I levels signifying PMI occurs during or immediately after prolonged, silent, stress-induced postoperative ischemia detected on continuous 12-lead ECG.
The definition of PMI
Postoperative MI was defined as a cTn-I level >3.1 ng/ml (12,17)after prolonged ST-segment deviation. This definition of MI conforms to the newly formulated European Society of Cardiology/American College of Cardiology consensus document redefining acute MI (18). Only 6 of the 12 patients with MI had either prolonged typical chest pain (5)or persistent new T wave inversion (1), thus fulfilling the traditional World Health Organization criteria for MI. Our ROC analysis of longest ischemia duration at different cutoff levels of cTn-I further confirmed that cTn-I levels between 3 and 3.5 ng/ml best defined PMI (Fig. 2).
Stress-induced postoperative ischemia
The majority (67%) of ischemic events, including those culminating in PMI, started at the end of surgery and emergence from anesthesia (Fig. 3), a time characterized by an increase in heart rate, blood pressure, sympathetic discharge and procoagulant activity (21). This finding is supported by two recent studies (12,22)in which the increase in troponin occurred mostly within 12 h to 24 h after surgery and implies that 36 h of monitoring may suffice to capture most PMIs. Also in accordance with previous studies, ischemia was denoted by ST depression, not elevation, and was preceded in all cases by an increase in heart rate. In all but one patient, ischemic ECG changes were transient, emphasizing the importance of continuous ST monitoring for its detection.
Pathophysiology of PMI
The mechanism underlying PMI is not known. It is assumed to resemble that of nonsurgical MI (23), that is, acute plaque rupture and coronary thrombosis caused by the abrupt increases in blood pressure, heart rate, coronary tone and platelet aggregability (24)and the decrease in fibrinolytic activity (25)occurring postoperatively. However, total coronary occlusion and coronary thrombosis are found in only 26% to 43% of patients early after either ST elevation or ST depression–type non-Q-wave infarction (26), compared with >90% coronary occlusion in patients with Q-wave infarction. None of our patients had ST elevation or Q-wave infarction. Coronary angiography was performed in only three of our patients with PMI within seven days of the infarction, and in each case it showed chronic, severe CAD without angiographically visible thrombus or ruptured plaques. In one patient the postoperative angiogram was identical to a previous one obtained six months earlier. Is it possible, therefore, that PMI in these patients occurred secondary to prolonged postoperative ischemia in the presence of severe, yet stable, CAD and not as a consequence of acute coronary occlusion?
Fuster et al. (27)proposed that, in contrast with the common thrombotic occlusion of a coronary artery after acute plaque disruption, thrombosis in a severe, but stable, coronary stenosis may result from a decrease in coronary blood flow and stasis. Tachycardia in the presence of fixed, but severe, coronary artery stenosis is one mechanism by which a significant decrease in coronary and myocardial blood flow occurs due to the shortening of diastolic time period (28). Vasoconstriction secondary to ischemia may further decrease coronary blood flow (29,30). Hence, if coronary thrombosis occurs in the postoperative setting, it may be the result of and not necessarily the cause of prolonged ischemia and PMI. Furthermore, diabetes mellitus and LVH were the independent preoperative predictors of PMI. Both conditions are associated with reduced coronary flow reserve and microvascular dysfunction (31,32), and both predict increased cardiac risk. To what extent reduced coronary flow reserve contributes to prolonged postoperative ischemia and PMI has yet to be determined.
There is only limited experimental or prospective clinical evidence for stress-induced MI. Myocardial adaptation or hibernation, rather than deterioration to infarction occurs after sustained (≤30 min) pacing-induced ischemia and norepinephrine infusion in animals with fixed coronary stenosis, provided that absolute myocardial blood flow is maintained (33). In contrast, prolonged (1 to 4 h) tachycardia-induced ischemia in the presence of a fixed, but critical, coronary stenosis leads to a progressive decline in subendocardial blood flow and diffuse subendocardial necrosis in anticoagulated dogs (34). One recent clinical study (15)showed reduced PMI and cardiac death rates from 34% to 3.4% by prophylactic bisoprolol treatment in high-risk patients undergoing major vascular surgery. This important report strongly supports our hypothesis that prolonged tachycardia and stress-induced ischemia underlie the evolution of PMI and that beta-blockade, by means of preventing prolonged postoperative ischemia, reduces PMI.
Although the mechanism of PMI remains unknown, our prospective study shows for the first time a strong temporal association between prolonged, silent, postoperative ischemia and PMI after major vascular surgery, suggesting that prolonged stress-induced ischemia may progress to MI. Whether early detection and treatment of prolonged postoperative ischemia reduce PMI has yet to be determined. Similarly, whether prolonged stress-induced ischemia may progress to non-Q-wave infarction in nonsurgical patients deserves further investigation.
☆ Supported by Chief Scientist, Ministry of Health of Israel (Grant no. 2828); Hebrew-University, Hadassah Hospital Combined Research Grant; Marquette Electronics, Inc., Office of Grants.
- coronary artery disease
- confidence interval
- creatine kinase
- cardiac troponin-I
- left ventricular hypertrophy
- myocardial infarction
- postoperative myocardial infarction
- receiver-operator characteristic curve
- Received June 20, 2000.
- Revision received January 30, 2001.
- Accepted February 13, 2001.
- American College of Cardiology
- Deedwania P.C,
- Carbajal E.V
- Pepine C.J,
- Sharaf B,
- Andrews T.C,
- et al.
- ACIP Investigators,
- Davies R.F,
- Goldberg A.D,
- Forman S,
- et al.
- Pepine C.J,
- Cohn P.F,
- Deedwania P.C,
- et al.
- Bodor G.S,
- Porter S,
- Ladt Y,
- Ladenson J.H
- Joint ESC/ACC Committee for the Redefinition of Myocardial Infarction
- Rosing D.R,
- Brakman P,
- Redwood D.R
- Nabel G.E,
- Selwyn A.P,
- Ganz P
- Sambuceti G,
- Marzilli M,
- Marraccini P,
- et al.
- Nahser P.J Jr.,
- Brown R.E,
- Oskarsson H,
- et al.
- Sobue T,
- Yokota M,
- Iwase M,
- Ishihara H
- Berman M,
- Fischman A.J,
- Southern J,
- et al.