Abstract
Objective: A major assumption in cardiovascular medicine is that Q-waves on the electrocardiogram indicate major myocardial tissue damage. The appearance of a new Q-wave has therefore been considered the most reliable criterion for diagnosis of perioperative myocardial infarction (PMI) in cardiac surgery. In a study, originally intended to evaluate troponin-T as a marker of PMI, analysis of our data aroused the need to address the reliability of Q-wave criteria for diagnosis of PMI. Methods: In 302 consecutive patients undergoing coronary surgery, Q-wave and other electrocardiogram (ECG) criteria were compared with biochemical markers of myocardial injury and the postoperative course. All ECGs were analysed by a cardiologist blinded to the biochemical analyses and the clinical course. Results: The incidence of positive Q-wave criteria was 8.1%. Combined biochemical (CK-MB≥70 μg/l) and Q-wave criteria were found in 1.0%. Patients with new Q-waves did not have CK-MB or troponin-T levels significantly different from those without Q-waves. More than 25% of the Q-waves were associated with plasma troponin-T below the reference level (<0.2 μg/l) on the fourth postoperative day. Q-wave criteria alone did not influence the postoperative course. In contrast, biochemical markers correlated with clinical outcome. Conclusions: The majority of Q-waves appearing after coronary surgery were not associated with major myocardial tissue damage, and according to troponin-T one-fourth of the Q-waves were not associated with myocardial necrosis. Furthermore, the appearance of Q-waves had little influence on short term clinical outcome. Therefore, the use of Q-wave criteria as the gold standard for diagnosis of PMI may have to be questioned.
Introduction
Perioperative myocardial infarction (PMI) is a major cause of mortality and morbidity after coronary surgery. In spite of this there is no consensus regarding diagnostic criteria and this accounts for the wide range in the reported incidence of PMI [1]. Enzyme criteria have been considered unreliable because of frequent unspecific elevation due to surgical tissue trauma, cardiopulmonary bypass, transient ischemia during aortic cross-clamping and retransfusion of mediastinal shed blood [2],[3],[4],[5]. As Q-waves on the electrocardiogram traditionally have been considered to signify major myocardial tissue damage [6], the appearance of a new Q-wave has gained acceptance as the most reliable criterion for diagnosis of PMI. Therefore, we chose electrocardiogram (ECG) criteria and Q-waves as the gold standard for evaluation of troponin-T as a marker of PMI. Preliminary analysis of our data, however, aroused the need to question the reliability of ECG criteria for diagnosis of PMI before further analysis of troponin-T and enzyme criteria could be undertaken. Consequently, the reliability of ECG, and Q-wave criteria in particular, for diagnosis of PMI will be addressed in the present analysis.
Patients
Three hundred and two consecutive patients undergoing routine coronary artery bypass surgery were included in the study. Patients undergoing redo-coronary operations or operations including valvular surgery or other procedures were excluded. Demographic data are presented in Table 1 .
Preoperative demographic data and intraoperative data for all patients included in the study (mean±SD)
Clinical management
On the day of surgery all patients were given their individual doses of betablockers and calcium-antagonists. After premedication with morphine hydrochloride and scopolamine, anesthesia was induced with thiopentone and fentanyl, and maintained with fentanyl and isoflurane. Pancuronium bromide was used for neuromuscular blockade.
Cardiopulmonary bypass was conducted with a membrane oxygenator and a roller pump generating non-pulsatile flow. Moderate hemodilution (hematocrit 20–25%) and moderate hypothermia (28–35°C) were employed. St. Thomas' cold crystalloid cardioplegic solution was used for myocardial protection. Shed mediastinal blood was routinely retransfused after surgery.
Methods
Study protocol
A 12-lead ECG was recorded using a Marquette electrocardiograph (Marquette Electronics Inc., Milwaukee, WI) the day before surgery, before anesthesia, on the morning after surgery (day 1) and on the fifth postoperative day (day 5). All ECGs were analysed by a cardiologist blinded to the results of the biochemical analyses and the clinical course. Furthermore, signs of sustained myocardial infarction according to the automatic interpretation program of the Marquette 12 SL system were also studied. To achieve reproducible ECGs the positions of the precordial leads were marked with permanent ink. The ECG findings were compared with the early postoperative course and biochemical markers of myocardial injury: aspartate aminotransferase (ASAT), creatine phosphokinase isoenzyme MB (CK-MB) and troponin-T. Blood samples were drawn preoperatively, 3 and 8 h after declamping the aorta, on the morning after surgery (day 1) and on the mornings of days 2–4. For the present analysis ASAT and CK-MB values on day 1 were used according to the clinical routine at our institution. Troponin-T is a cardiospecific protein and due to its kinetics myocardial necrosis is characterized by markedly elevated levels of troponin-T for several days, as demonstrated in cardiology practice and after cardiac surgery [7],[8],[9]. Therefore, troponin-T results obtained on the 4th postoperative day were used for evaluation of ECG criteria. A normal troponin-T level (reference value<0.2 μg/l) was considered to suggest that no (or only minimal) myocardial tissue necrosis had been sustained. Furthermore, as different release patterns of biochemical markers could be expected in patients with and without permanent myocardial damage, plasma kinetics of CK-MB and troponin-T were compared in patients with and without positive Q-wave criteria [10],[11],[12].
Calculations and definitions
A new Q-wave was defined as the appearance of a Q-wave≥40 ms in at least two adjacent leads [13]. The Q-waves were measured manually on the averaged ECGs registered with the Marquette equipment.
The Marquette-criteria were determined by the Marquette Electronics 12 SL analysis program (statement of validation and accuracy, MEI, 1991). These criteria consist of algorithms combining mainly Q-wave depth and amplitude, and R-wave amplitude and progression, and secondarily in combination with these, T-wave abnormalities. These criteria have been validated with reference to clinical information and coronary angiography (statement of validation and accuracy, MEI, 1991). However, they were not developed for detection of myocardial infarction in cardiac surgical practice.
Biochemical markers were considered suggestive of PMI if routine criteria currently employed at our institution were fulfilled (CK-MB≥70 μg/l on the first postoperative day). However, if CK-MB data were incomplete an ASAT day 1 value≥3.0 μkat/l associated with a troponin-T day 4 value≥2.0 μg/l was also considered suggestive of PMI (Table 4).
Biochemical analyses
Troponin-T was measured with reagents from Boehringer Mannheim, using the ES 300 equipment. ASAT activity was measured using the Hitachi 717 instrument with Boehringer Mannheim reagents according to IFCC specifications. CK-MB mass was measured using the Abbot CK-MB stat method employing the Imx analytical instrument.
Statistical methods
Analysis of variance (ANOVA), including post-hoc comparisons with the Tukey honest significant difference test, was employed for statistical analysis involving multiple groups (continuous data). Chi-square analysis was used for categorical data. Student's t-test was used when only data from two groups were available for comparison. Statistical significance was defined as P<0.05. Data are presented as mean±standard deviation (SD). Statistical analyses were performed using a computerized statistical package (Statistica 5.1, StatSoft Inc., Tulsa, AR).
Ethical aspects
The study was approved by the Ethical Committee for Medical Research at the University Hospital in Linköping. Informed consent was obtained from each patient.
Results
Biochemical tests
The average values for all patients included in the study were for ASAT day 1: 1.82±1.13 μkat/l, CK-MB day 1: 37.4±46.7 μg/l and for troponin-T day 4: 0.89±1.65 μg/l.
Biochemical markers suggestive of PMI were found in 30 patients (10%). Of these 27 patients had CK-MB≥70 μg/l on the first postoperative day and 19 patients had an ASAT day 1 value≥3.0 μkat/l associated with a troponin -T day 4≥2.0 μg/l.
Electrocardiogram
Data from three patients could not be analysed as they died during the first postoperative day. A further three patients died during the study period but they were included in the analysis as ECGs and laboratory data from the first postoperative day were available. Five patients had left bundle branch block (LBBB) preoperatively and their ECGs were therefore excluded from further analysis. Seven patients had preoperative right bundle branch block (RBBB), but they were not excluded as the diagnosis of myocardial infarction based on a new Q-wave can be established in RBBB.
Postoperatively seven patients had new LBBB and 27 had new RBBB. These two groups were analysed separately and compared with patients without bundle branch block (BBB) or Q-wave criteria.
Q-waves
Of the 260 patients with an ECG suitable for analysis, 42 patients had a Q-wave present before surgery. A new Q-wave was recorded on day 1 in 19 patients. When ECG was recorded on day 5 the Q-wave had disappeared in three patients, whereas a further three patients had developed a new Q-wave. Together twenty-two patients (8.1%) developed a new Q-wave during the study period. Combined Q-wave criteria and biochemical markers suggestive of PMI were found in three patients (1.0%).
The average values for biochemical markers did not differ in patients with and without positive Q-wave criteria (Table 2 ). Plasma kinetics of troponin-T and CK-MB for patients with and without positive Q-wave criteria (day 5) are presented in Fig. 1Fig. 2 . There were no statistically significant differences in plasma levels at any time during the study period. Average peak value for CK-MB and troponin-T was recorded 8 h after declamping in patients with new Q-waves whereas average peak value for CK-MB and troponin-T in patients without Q-wave criteria was recorded 3 h after declamping. In the Q-wave group 12 patients had their CK-MB peak at 3 h after declamping (54.5%), six patients at 8 h after declamping and four patients on the first postoperative morning (18.2%). This distribution did not differ significantly from what was found in patients without new Q-waves. In the latter group 52.4% had their peak value recorded 3 h after declamping and 10.7% on the first postoperative morning or later.
Enzymes and troponin-T in patients with (Pos) and without positive (Neg) electrocardiographic Q-wave or Marquette criteria after coronary surgery
Plasma kinetics of troponin-T in patients with positive Q-wave criteria day 5 (dotted line+(▽)) and in patients without Q-wave criteria (solid line+(▴)). The error bars indicate SD. There were no statistically significant differences between the groups at any time during the study period.
Plasma kinetics of CK-MB in patients with positive Q-wave criteria day 5 (dotted line+(▽)) and in patients without Q-wave criteria (solid line+(▴)). The error bars indicate SD. There were no statistically significant differences between the groups at any time during the study period.
According to troponin-T analyses five out of 19 patients with new persistent Q-waves (day 5) had no signs of myocardial tissue necrosis (troponin-T day 4<0.2 μg/l).
Marquette criteria
Preoperatively 91 patients had electrocardiographic signs of previous myocardial infarction, according to the Marquette 12 SL analysis program. After surgery ECG changes fulfilling these criteria appeared in 42 patients on day 1, and in 40 patients on day 5. Combined Marquette criteria and biochemical markers suggestive of PMI were found in ten patients (3.3%).
The average values for ASAT day 1, CK-MB day 1 and troponin-T day 4 were significantly elevated compared with patients not fulfilling the Marquette criteria (Table 2). However, ten out of the 40 patients with positive Marquette criteria (day 5) had no signs of myocardial tissue necrosis according to troponin-T analyses (troponin-T day 4<0.2 μg/l).
Bundle branch block (BBB)
Of the 265 patients included in the analysis there were seven patients with a new LBBB and 27 with a new RBBB after surgery. The average values for biochemical markers in patients with a new LBBB or RBBB were not significantly different from those without new BBB or Q-wave (Table 3 ). Sixteen percent of the patients with new RBBB had no signs of myocardial tissue necrosis (troponin-T day 4<0.2 μg/l). None of the patients with a new LBBB had troponin-T values day 4 below the reference level.
Enzymes and troponin-T in patients with and without a new (BBB after coronary surgery
Postoperative course (Table 4)
Postoperative course in patients with and without criteria suggesting perioperative myocardial infarction
Patients with positive Q-wave criteria or BBB alone had an essentially unaffected postoperative course. In contrast, patients with biochemical markers suggestive of PMI had an impaired postoperative course. Mortality was significantly higher and pharmacological and mechanical circulatory support was more often required in these patients. Furthermore, ventilator treatment and intensive care unit stay was significantly prolonged.
Discussion
The diagnosis of acute myocardial infarction in clinical cardiology is usually based on a history of chest pain, elevation of cardiac enzymes and ECG criteria. As all these diagnostic modalities are influenced by cardiac surgery, the difficulties to establish a PMI diagnosis are evident [1]. However, it is a major assumption in cardiovascular medicine that pathological Q-waves on the electrocardiogram indicate major myocardial tissue damage. The appearance of such Q-waves have therefore gained acceptance as the most reliable criteria for diagnosis of PMI in cardiac surgery.
Enzyme criteria have been considered unreliable because of frequent unspecific elevation due to surgical tissue trauma, cardiopulmonary bypass, transient ischemia during aortic cross-clamping and retransfusion of mediastinal shed blood [2],[3],[4],[5]. On the other hand, this `background noise' may be expected to influence the enzyme levels of patients with or without PMI to a similar degree. As positive Q-wave criteria are assumed to represent moderate to large (although not necessarily transmural) infarcts [6], clearly elevated levels of biochemical markers were anticipated. However, we found no significant differences in the magnitude of CK-MB (or ASAT or troponin-T) elevations between patients with and without positive Q-wave criteria (Fig. 1Fig. 2, Table 2). Patients with permanent myocardial damage may also be expected to have delayed release patterns of biochemical markers from the heart compared with patients having sustained no or minimal myocardial damage [10],[11],[12]. However, only minor differences were found between patients with and without positive Q-wave criteria (Fig. 1Fig. 2) and the majority of patients with Q-waves had their peak values of CK-MB recorded 3 h after declamping. Furthermore, no significant differences in the levels of CK-MB or troponin-T between the groups were found at any time during the study period.
The results of this study also raises question marks regarding the clinical significance of positive Q-wave criteria as the appearance of Q-waves alone had no significant influence on short-term clinical outcome. In contrast, the results suggest that biochemical markers may reflect myocardial damage better than previously assumed, as they correlated with clinical outcome (Table 4).
Occasional reports have previously questioned the validity of Q-waves for diagnosis of PMI, as Q-waves have been demonstrated to appear in a variety of conditions unrelated to myocardial infarct [14],[15],[16]. However, these papers have been based on observations from a limited number of cases. Recently it was demonstrated in a larger cohort of patients (n=194) undergoing coronary surgery that Q-wave infarction identified in the postoperative period was a weak end point with limited prognostic significance. Although that study was retrospective, the data retrieved clearly showed that the appearance of Q-waves correlated poorly with postoperative catheterization data (including wall motion abnormalities), enzymes and clinical outcome [17]. The results of our prospective study are consistent with these findings regarding the relationship between Q-waves and biochemical markers of myocardial injury, and the impact of Q-waves on short-term clinical outcome.
As pathological Q-waves undoubtedly may indicate major myocardial infarcts after cardiac surgery, studies on high risk patients with a higher incidence of infarcts and studies employing less strict Q-wave criteria and incorporating clinical judgement may be expected to find a better agreement between ECG and biochemical markers [18],[19]. In the present study, however, the interpretations were made by an experienced cardiologist blinded to the biochemical data and the clinical course. Strict adherence was kept to the criteria and no subjective modification was allowed.
Thus, it seems that Q-wave criteria alone in the majority of cases do not indicate significant myocardial damage and hence their appearances are of limited clinical significance after coronary artery bypass graft (CABG) procedures. It can be argued that the appearance of Q-waves after cardiac surgery also could indicate minor myocardial infarcts. However, as troponin-T is mainly structurally bound, myocardial necrosis characteristically leads to a prolonged elevation of plasma troponin-T lasting up to a week or longer [1],[8],[9],[20],[21],[22]. Therefore, the results suggest that more than 25% of the Q-waves were associated with no or minimal myocardial tissue necrosis. To explain these observations, the definition of Q-wave criteria and the electrophysiology behind the development of Q-waves deserve consideration.
The Q-wave criteria in the present study did not include depth of the Q-wave [13]. However, we found that all patients complying with these criteria also had Q-amplitudes greater than one-third of R [23]. Therefore, the lack of specificity cannot be attributed to the choice of inadequate Q-wave criteria. Abnormal Q-waves reflect the loss of electrically active myocardium in the area represented by the leads where they appear. As the different parts of the left ventricle are not equally represented in the ECG some myocardial infarcts may not cause Q-waves while others even smaller will. Accordingly, 40–50% of acute myocardial infarcts in cardiology practice cannot be diagnosed by conventional ECG criteria [24],[25]. The sequences of activation of the ventricles explain why parts of the left ventricle activated first are more likely to exhibit Q-waves in acute myocardial infarction, while a myocardial injury in an area that is activated later will result in slurring or notching of the QRS-complex or a loss of R amplitude [24]. A false positive Q-wave can be anticipated in cases where a small R-wave seen preoperatively disappears leaving a pathological Q-wave although the myocardial injury may be very limited. As 91 of the patients had signs of previous myocardial infarct in the preoperative ECG, it may have influenced the outcome of this study. In the literature `pseudo-Q waves', unmasking of pre-existent Q-waves by improved function of the contralateral ventricular segment and transient abnormal Q-waves due to temporary alterations in ventricular polarization have been described [2],[15],[16]. These types of electrophysiological phenomena could explain some uncertainties concerning Q-waves.
In recent years there has been a rapid development of computerized interpretation programs for diagnosis of myocardial infarction, with improved diagnostic sensitivity compared with older criteria. The Marquette 12 SL is such a system, but its algorithms for detection of sustained myocardial infarction were developed for general cardiological use. In cardiac surgical practice Marquette criteria have not received widespread use since unspecific ECG changes are common. Nevertheless, our results showed that Marquette criteria correlated better with biochemical markers than Q-wave criteria alone, and the specificity was not substantially lower as the proportion of patients without myocardial necrosis according to troponin-T analyses was similar. The development of specific algorithms for computerized detection of perioperative myocardial infarction would be desirable.
Development of a new BBB after cardiac surgery may indicate a myocardial infarct [18]. However, several benign causes of new BBB have been described, such as effects of cold crystalloid cardioplegia, intraoperative kinking of the septum and transient perioperative ischemia [26],[27]. This is in keeping with the present study. The appearance of a BBB alone did not correlate with either biochemical markers or clinical outcome.
The results of this study should not be interpreted as though Q-waves and ECG are of little importance in cardiac surgery. Obviously Q-wave criteria should be interpreted cautiously and preferably together with biochemical and clinical data. In selective patients with concomitant elevation of biochemical markers electrocardiographic changes may contribute to localizing infarcts. Furthermore, electrocardiographic monitoring remains an invaluable tool for detection of myocardial ischemia and myocardial arrhythmias. Recently published data suggest that intraoperative ST segment changes during CABG procedures may have substantial prognostic implications [28].
To conclude, the present study suggests that the role of ECG in general, and the Q-wave in particular, for diagnosis of PMI after cardiac surgery has to be re-evaluated. Obviously, the diagnosis of PMI cannot be based on ECG criteria alone. However, as true Q-wave infarcts undoubtedly occur after cardiac surgery, ECG changes can with biochemical markers and clinical findings contribute to the interpretation.
