Author + information
- Received November 21, 2011
- Revision received February 3, 2012
- Accepted February 14, 2012
- Published online October 1, 2012.
- Anca Florian, MD⁎,
- Massimo Slavich, MD⁎,
- Pier Giorgio Masci, MD†,
- Stefan Janssens, MD, PhD‡ and
- Jan Bogaert, MD, PhD⁎,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Jan Bogaert, Department of Radiology, UZ Leuven, Herestraat 49, B-3000 Leuven, Belgium
Objectives The aim of this study was to evaluate the evolution in Q-wave expression during the first 5 years after a primary, successfully reperfused ST-segment elevation myocardial infarction (MI), using cardiac magnetic resonance (CMR) for infarct location, and to depict changes in infarct size and left ventricular remodeling over time.
Background In the absence of QRS confounders, abnormal Q waves are usually diagnostic of myocardial necrosis. It is hypothesized that Q-wave regression after MI could be related to smaller infarct sizes. Late gadolinium enhancement accurately depicts MI of any age.
Methods Forty-six MI patients underwent electrocardiography and CMR at 1 week (baseline), 4 months, 1 year, and 5 years post-infarction. Conventional CMR parameters were analyzed, and infarct presence, location, and size were assessed using late gadolinium enhancement CMR. Infarct locations were anterior or nonanterior (inferior and/or lateral), using late gadolinium enhancement CMR as a reference. For each time point, patients were classified as having a diagnostic/nondiagnostic electrocardiogram (ECG) using the European Society of Cardiology/American College of Cardiology Foundation/American Heart Association/World Heart Federation consensus criteria for previous Q-wave infarct.
Results At baseline, 11 patients (23%) did not meet the criteria for Q-wave MI. Non–Q-wave infarcts were significantly smaller than Q-wave infarcts (p < 0.0001). All anterior Q-wave infarcts (n = 17) were correctly localized, whereas in 7 of 19 nonanterior Q-wave infarcts, the location or extent of the infarct was misjudged by electrocardiography. At 4-month/1-year follow-up, in 10 patients (3 anterior/7 nonanterior), the ECG became nondiagnostic. The ECG remained nondiagnostic at 5-year follow-up. A cutoff infarct size of 6.2% at 1 year yielded a sensitivity of 89% and a specificity of 74% to predict the presence or absence of Q waves.
Conclusions The incidence of nondiagnostic ECGs for previous MI using the current European Society of Cardiology/American College of Cardiology Foundation/American Heart Association/World Heart Federation criteria is substantial and increases with time post-infarction from 23% immediately post-infarction to 44% at 5-year follow-up.
In the absence of QRS confounders, abnormal Q waves on the surface electrocardiogram (ECG) are usually diagnostic of myocardial necrosis. They develop in the first hours after the onset of an ST-segment elevation myocardial infarction (STEMI) and persist for a variable amount of time, often indefinitely (1). Based on the depth and width and ratio to an R-wave, several validated criteria for abnormal Q waves are available: European Society of Cardiology/American College of Cardiology Foundation/American Heart Association/World Heart Federation (ESC/ACCF/AHA/WHF) consensus criteria, Minnesota code, Novacode, and WHO MONICA (2). Regression of Q waves after myocardial infarction (MI) is related to lower left ventricular (LV) end-diastolic pressures, higher ejection fraction, and reduced risk of LV aneurysm formation and congestive heart failure, suggesting that Q-wave loss may be related to smaller infarct sizes. Nowadays, in the “era of reperfusion therapy” and primary percutaneous coronary interventions (PCIs), the disappearance of Q waves occurs even more frequently (3).
Late gadolinium enhancement (LGE) cardiac magnetic resonance (CMR) is a well-validated tool for accurate and reproducible visualization of irreversible myocardial damage in the acute and chronic settings of MI (4). Several studies have shown that the presence of diagnostic Q waves is primarily determined by MI size rather than its transmural extent (5–9). The larger the endocardial extent of MI is, more likely it is that the ECG will be diagnostic for Q-wave MI (10). Moreover, ECG-derived estimates available for infarct size estimation correlate only modestly with those of LGE CMR (11).
The objective of the present study was to assess the electrocardiographic changes in Q-wave expression in the first 5-year period after a first reperfused STEMI and to relate electrocardiographic findings to parameters gauging infarct extent and LV remodeling as assessed by CMR.
From a double-blind, randomized, controlled study that investigated the effect of autologous bone marrow–derived stem cell transfer on LV remodeling after STEMI treated by primary PCI (within 12 h from symptom onset), we retrospectively identified 48 patients who met the following criteria: 1) first MI; 2) no confounders for Q-wave analysis on the ECG (left or right bundle branch block, LV hypertrophy with strain or paced rhythms); and 3) both ECG tracings and CMR studies available in the first week (baseline) and at 4 months, 1 year, and 5 years after the acute event (12). Two patients were excluded: 1 who had experienced in-stent thrombosis 11 months after MI and 1 who had an MI in a different coronary territory during follow-up. The local ethics review board approved the protocol, and written informed consent was obtained from each patient.
CMR data acquisition
CMR studies were performed on a 1.5-T unit (Intera-CV, Philips, Best, the Netherlands) using commercially available cardiac software, electrocardiographic triggering, and cardiac-dedicated surface coils. CMR included cine imaging, T2-weighted imaging, and LGE-CMR, as previously described in detail (12).
CMR data analysis
CMR studies were analyzed blinded to the clinical and electrocardiographic data. Functional parameters included LV volumes at end-diastole and end-systole, ejection fraction, and myocardial mass. The area at risk (AAR) was determined by T2-weighted imaging. LGE-CMR was used to quantify microvascular obstruction and infarct mass and its relative extent (normalized LV mass). The salvage index was calculated as the difference between AAR and baseline infarct size normalized to AAR. A 5-grade score (0 = no LGE; 1 = 0 to 25%; 2 = 26% to 50%; 3 = 51% to 75%; and 4 = 76% to 100% LGE) was used to express infarct transmurality using the 17-segment model as recommended by the American Heart Association. Per patient, a transmurality score was obtained by adding the segmental grades (13). A transmural infarct was defined as a transmural score of 4 in at least 1 segment.
Standard 12-lead ECGs obtained at the time of CMR were recorded at a speed of 25 mm/s and a voltage of 10 mm/mV. Studies were randomly analyzed by 2 cardiologists blinded to clinical and CMR data. Any disagreement was resolved by a consensus reading. The ESC/ACCF/AHA/WHF criteria for previous MI were used to assess patterns of necrosis on the ECG as follows: 1) Q waves were considered pathological (Q) if ≥0.02 s (or QS complex) in leads V2,V3, and ≥0.03 s and 0.1 mV deep (or QS complex) in leads I, aVL, V6, V4 to V6, II, III, aVF; and 2) R waves were considered pathological if ≥0.04 s and R/S ≥1 in leads V1 and/or V2 with a concordant positive T-wave in the absence of right-axis deviation ≥100° (1,2,5). In the anterior precordial leads, an R-wave was considered the initial positive deflection of >0.1-mV amplitude. The presence, location, and number of Q waves were noted for each tracing. To better characterize the relationship between the number of pathological Q waves and infarct size on CMR, isolated Q waves in any of the above-mentioned lead groupings were counted in the total number of Q waves of a patient.
Localization of MI
Infarct location was attributed to the LV walls, as determined by LGE-CMR, and not to a specific coronary artery territory. Thus, anterior MI was defined if LGE was present in at least 1 of the following segments: basal anteroseptal, midanterior, midanteroseptal, or apical anterior (i.e., anterior MI group). Inferior MI was defined if LGE involved at least 1 of the basal and midventricular inferior segments, whereas lateral MI was considered if there was LGE in at least 1 of the lateral segments (14). Inferior and lateral MIs were considered as the nonanterior MI group. On an ECG, anterior MI was defined if Q waves were present in V1 to V6 leads, inferior MI if Q waves were present in any 2 inferior leads (II, III, aVF), and lateral MI if Q waves were present in any 2 of the leads I, aVL, or V6, or if pathological R waves were present in leads V1 to V2. Pathological R waves in leads V1 and/or V2 were considered lateral Q-wave equivalents. The presence of a Q wave in aVL ± I and in V2 to V3 but not in V6 was considered anterior MI (15).
Continuous variables were expressed as mean ± SD. Skewed variables were expressed as median and interquartile range. Categorical variables were expressed as frequency with percentage. Student t test was used to compare baseline patient characteristics expressed as continuous variables. Repeated-measures analysis of variance with a post hoc Bonferroni test was used to assess timely changes in CMR parameters in and between patients groups. Nonparametric tests were used for not normally distributed variables (i.e., Mann-Whitney U test and Friedman test for repeated measurements). For the time changes in relative infarct size in anterior/nonanterior groups, where sphericity was violated (Mauchly test), analysis of variance with repeated measures with a Greenhouse-Geisser correction was used. The chi-square test was used to compare noncontinuous variables, expressed as proportions. Pearson correlation (r) was used to assess the relationship between infarct extent on CMR and the number of Q waves. Receiver-operating characteristic curve analysis was used to determine the best cutoff value of infarct size to predict the presence of a diagnostic ECG. The cutoff was identified as the point on receiver-operating characteristic curve closest to the upper left corner. Statistical analysis was performed using SPSS software for Windows (version 18, SPSS Inc., Chicago, Illinois). A p value <0.05 was considered statistically significant.
Baseline CMR studies and ECG tracings were obtained at day 4 (range 3 to 5 days) and at day 7 (range 5 to 9 days) post-PCI, respectively. Patients were equally distributed between anterior and nonanterior MIs (Table 1). In 36 of 46 patients (78%), Q waves were present (Fig. 1). Patients with a nondiagnostic ECG (i.e., 6 anterior, 4 nonanterior) had a significantly smaller AAR and infarct size and smaller LV volumes and mass than patients with Q waves (Table 2). Q waves developed in none of these patients on their ECG at follow-up. In patients with a diagnostic ECG, the infarct was anterior (n = 17), inferior (n = 14), lateral (n = 1), and mixed inferior/lateral (n = 4) on LGE-CMR. Although the ECG correctly located all anterior MIs, in 4 of 14 inferior MIs, Q waves were also present in the lateral leads. The patient with the lateral MI had an ECG diagnostic of inferior MI. Finally, an ECG/LGE-CMR match was found in 2 of 4 patients with a mixed inferior/lateral MI (Fig. 2).
Anterior infarcts exhibited a significantly larger infarct size than nonanterior infarcts (Table 3). A transmural MI was present in 15 and 17 of anterior and nonanterior infarcts, respectively. The median number of Q waves was 3 (range 0 to 7) and 3 (range 0 to 5) for anterior and nonanterior infarcts, respectively (p = 0.5). No correlation was found between number of Q waves and infarct size.
The same distribution of LGE as at baseline, except in 2 patients with a inferior/lateral MI in whom LGE was limited to the inferior wall at 1 year. Both patients showed subendocardial LGE at baseline, and the lateral wall was only partially involved. Infarct size significantly decreased at follow-up (i.e., from 20 ± 12 g at baseline to 12 ± 9 g, 10 ± 8 g, and 9 ± 7 g at 4 months, 1 year, and 5 years, respectively [p = 0.001 for trend]). Anterior infarcts remained significantly larger than nonanterior infarcts (Table 3). Transmural MI was found at 5 years in 14 of 15 anterior and 12 of 17 nonanterior infarcts with a baseline transmural MI.
Ten patients (21%) developed nondiagnostic ECG, 6 patients at 4 months (3 anterior, 3 nonanterior), and 4 patients at 1 year (all nonanterior) (Figs. 1, 3, and 4).⇓ Three patients with an inferior/lateral infarct on the baseline ECG demonstrated an inferior infarct on the 1-year ECG. In 2 of them, baseline CMR showed a pure inferior infarct. Two inferior infarcts on the baseline CMR/ECG appeared on the 1-year ECG as an inferior/lateral infarct and lateral infarct, respectively. The number of Q waves did not differ significantly over time. In anterior infarcts, a moderate correlation was found between relative infarct size and number of Q waves (i.e., r = 0.56 [p = 0.006] at 4 months; r = 0.58 [p = 0.005] at 1 year, and r = 0.56 [p = 0.019] at 5 years).
CMR versus ECG for previous MI detection
The frequency of patients with nondiagnostic ECGs increased from 22% at baseline to 43% at 5 years. By receiver-operating characteristic curve analysis, relative infarct size predicted a diagnostic ECG with an area under the curve of 0.76 (95% confidence interval: 0.61 to 0.91) at baseline, which increased to 0.84 (95% confidence interval: 0.71 to 0.96) at 5 years. A cutoff value with 6.2% of relative infarct size at 1 year yielded the highest sensitivity (89%) and specificity (74%), with an area under the curve of 0.85 (95% confidence interval, 0.75 to 0.97) (Fig. 5). In nonanterior infarct patients having a nondiagnostic ECG over time, the baseline relative infarct size was significantly smaller than in patients with a diagnostic ECG (Fig. 6). The degree of infarct shrinkage was similar among groups. Although patients with a nondiagnostic ECG at 5 years had higher ejection fractions and lower end-systolic volumes, no statistical significance was reached, except for end-systolic volume at 4 months (Fig. 7).
Although the 12-lead surface ECG is the frontline tool in the diagnosis of acute and healed MI (2), the present study shows that the current ESC/ACCF/AHA/WHF ECG consensus criteria for previous MI frequently fail to depict MI in patients with a first-time STEMI treated by primary PCI. Moreover, the number of nondiagnostic ECGs increases over time after the acute event. In particular, small infarcts and infarcts in nonanterior locations at baseline are those more commonly associated with nondiagnostic ECGs at mid- and long-term follow-up.
In the absence of QRS confounders, abnormal Q waves are usually diagnostic of myocardial necrosis. Using LGE-CMR as in vivo validation technique, the presence or absence of Q waves on the ECG is primarily determined by total infarct size (i.e., endocardial infarct extent and not by transmural extent) (5–10,16,17). Moreover, ECG-derived estimates of infarct size correlate only modestly with those with LGE-CMR, especially in lateral infarcts, which are often electrically silent (11,18,19).
The novelty of our approach is that patients with MI were studied at 4 time points post-infarction, providing insight into short-, intermediate-, and long-term infarct and ventricular remodeling. Although all patients had a well-documented acute STEMI, only 36 of 46 of the analyzable patients (78%) had a diagnostic ECG early post-infarction. Similar findings were reported by Engblom et al. (10), using the Minnesota ECG criteria for MI detection, with nondiagnostic ECGs in 11 of 29 patients 1 week after primary PCI. In particular, there is a risk that smaller infarcts remain undiagnosed, most likely because they generate insufficiently large Q waves to be transmitted to the body surface. An ECG is excellent for locating anterior infarcts, but it not infrequently fails to accurately locate nonanterior infarcts. Lateral infarcts or inferior infarcts extending to the lateral wall are difficult to depict on an ECG (20). As suggested by Rovai et al. (18), several factors may contribute to a lower sensitivity of an ECG to detect infarcts in this part of the ventricle (21). Additionally, the number of Q waves (or Q-wave equivalents), as a measure of infarct severity, correlates poorly in the acute phase with LGE measures (such as infarct size and transmurality), whereas a moderate positive correlation is found at long-term phases for anterior infarcts. Ibrahim et al. (22) showed significant changes in LGE extent in the first week post-infarction, which may explain the lack of agreement between ECG and CMR findings early post-infarction.
Remarkably, the number of patients with nondiagnostic ECGs for previous MI doubled at 5 years post-infarction. A new nondiagnostic ECG always occurred in the first year post-infarction. Taking into account the significant reduction in infarct size in the first months post-infarction (as measured by LGE-CMR), infarct size may drop below a critical threshold to yield a diagnostic ECG. In particular, most of the new nondiagnostic ECGs at follow-up occurred in the nonanterior group, and this is likely explained by the fact that nonanterior infarcts had smaller baseline infarct size than anterior infarcts; the diagnostic performance of ECG is also reduced when depicting infarcts in this region of the ventricle.
Our study results indicate that a cutoff (relative) infarct size of 6.2% (at 1 year) yields good sensitivity and moderate specificity to predict the presence/absence of Q waves. This cutoff is considerably lower than the cutoff value (i.e., 17%) previously reported by Kaandorp et al. (6), suggesting that an ECG is able to depict smaller infarcts. This discrepancy can be explained by a difference in treatment strategies because in the study of Kaandorp et al. (6), half of the patients were treated conservatively. Accordingly, the relative infarct size was larger in their study (11 ± 9% vs. 8 ± 7% at 4 months in our study).
Putting our findings in a clinical perspective, it is important to realize that normalization of the ECG post-infarction is frequent (23), even in patients with a well-documented acute STEMI, and confirms what was already suggested by Cox more than 4 decades ago (24). However, with the advent of LGE-CMR as an in vivo validation technique, it has become clear that it concerns a pseudo- and not a true normalization because irreversible myocardial damage continues to persist at late follow-up. Because these electrically silent infarcts have a prognosis similar to that of overt ones, use of more accurate techniques such as LGE-CMR may be indicated to depict myocardial damage in patients with suspected previous MI (25–27).
The small number of patients is an important limitation of the study. Nonetheless, the design is unique because all patients were investigated at 4 time points post-infarction using LGE-CMR, the current noninvasive reference of MI detection, thereby providing distinctive data regarding the MI evolution over a period of 5 years. Second, we considered only the criteria for previous MI, without taking into consideration the pre-PCI ECG tracings and other parameters as ST-segment shifts in the acute phase. Thus, no inferences can be made regarding the influence of ST-segment shifts post-PCI on the further Q-wave expression. Third, patients with ECG confounders such as the presence of bundle branch blocks were excluded from the analysis. We can only assume that if these patients had been included, the accuracy of ECG in diagnosing and localizing MI would have been less. Last, a qualitative approach rather than a scoring system was used to evaluate the Q-wave changes over time post-infarction. We opted for this approach because it is better suited to daily clinical practice. Moreover, in the “era of reperfusion therapy,” these time-consuming scoring systems, like the Sylvester score, have proved to be less accurate compared with CMR-LGE (11).
The electrocardiographic appearance of Q waves in patients with a previous MI is mainly determined by infarct size. The ESC/ACCF/AHA/WHF consensus criteria frequently fail to depict small size STEMIs in nonanterior locations. Post-infarction electrical Q-wave remodeling parallels infarct remodeling, resulting in nearly double the nondiagnostic ECGs at 5 years post-infarction.
This study was funded in part by a grant from Research Foundation Flanders (G.0613.09). All authors have reported that they have no relationships relevant to the contents of this paper to disclose. Drs. Florian and Slavich contributed equally to this paper.
- Abbreviations and Acronyms
- area at risk
- cardiac magnetic resonance
- European Society of Cardiology/American College of Cardiology Foundation/American Heart Association/World Heart Federation
- late gadolinium enhancement
- left ventricular
- myocardial infarction
- percutaneous coronary intervention
- Received November 21, 2011.
- Revision received February 3, 2012.
- Accepted February 14, 2012.
- American College of Cardiology Foundation
- Surawicz B.,
- Knilans T.K.
- Thygesen K.,
- Alpert J.S.,
- White H.D.,
- Joint ESC/ACCF/AHA/WHF Task Force for the Redefinition of Myocardial Infarction
- Kim H.W.,
- Farzaneh-Far A.,
- Kim R.J.
- Moon J.C.,
- De Arenaza D.P.,
- Elkington A.G.,
- et al.
- Geerse D.A.,
- Wu K.C.,
- Gorgels A.P.,
- Zimmet J.,
- Wagner G.S.,
- Miller J.M.
- Cerqueira M.D.,
- Weissman N.J.,
- Dilsizian V.,
- et al.,
- American Heart Association Writing Group on Myocardial Segmentation and Registration for Cardiac Imaging
- Ortiz-Pérez J.T.,
- Rodríguez J.,
- Meyers S.N.,
- Lee D.C.,
- Davidson C.,
- Wu E.
- Bayés de Luna A.,
- Wagner G.,
- Birnbaum Y.,
- et al.,
- International Society for Holter and Noninvasive Electrocardiography
- Smalling R.W.
- Barbier C.E.,
- Bjerner T.,
- Johansson L.,
- Lind L.,
- Ahlström H.
- Krittayaphong R.,
- Maneesai A.,
- Chaithiraphan V.,
- Saiviroonporn P.,
- Chaipet O.,
- Udompunturak S.
- Kwong R.Y.,
- Sattar H.,
- Wu H.,
- et al.