Author + information
- Received August 10, 1998
- Revision received September 21, 1999
- Accepted November 15, 1999
- Published online March 1, 2000.
- Akbar Shah, MDa,* (, )
- Galen S Wagner, MDa,
- Christopher B Granger, MD, FACCa,
- Christopher M O’Connor, MD, FACCa,
- Cynthia L Green, MSa,
- Kathleen M Trollinger, RNa,
- Robert M Califf, MD, FACCa and
- Mitchell W Krucoff, MD, FACCa
- ↵*Reprint requests and correspondence: Dr. Akbar Shah, 1380 E. Stroop Road, Cardiology South, Inc., Kettering, Ohio 45429
To compare the prognostic significance of reperfusion assessment by Thrombolysis in Myocardial Infarction (TIMI) flow grade in the infarct related artery and ST-segment resolution analysis, by correlating with clinical outcomes in patients with acute myocardial infarction (AMI).
Angiographic assessment, based on epicardial coronary anatomy, has been considered the “gold standard” for reperfusion. The electrocardiogram (ECG) monitoring provides a noninvasive, real-time physiologic marker of cellular reperfusion and may better predict clinical outcomes.
Two hundred fifty-eight AMI patients from the Thrombolytics and Myocardia Infarction phase 7 and Global Utilization of Streptokinase tPA for Occluded coronary arteries phase 1 trials were stratified based on blinded, simultaneous reperfusion assessment on the acute angiogram (divided into TIMI grades 0 & 1, TIMI grade 2 and TIMI grade 3) and ST-segment resolution analysis (divided into: <50% ST-segment elevation resolution or reelevation and ≥50% ST-segment elevation resolution). In-hospital mortality, congestive heart failure (CHF) and combined mortality or CHF were compared to determine the prognostic significance of reperfusion assessment by each modality using chi-square and Fisher’s Exact tests for univariable correlation and logistic regression analysis for univariable and multivariable prediction models.
By logistic regression analysis, ST-segment resolution patterns were an independent predictor of the combined outcome of mortality or CHF (p = 0.024), whereas TIMI flow grade was not (p = 0.693). Among the patients determined to have failed reperfusion by TIMI flow grade assessment (TIMI flow grade 0 & 1), the ST-segment resolution of ≥50% identified a subgroup with relatively benign outcomes with the incidence of the combined end point of mortality or CHF 17.2% versus 37.2% in those without ST-segment resolution (p = 0.06).
Continuous 12-lead ECG monitoring can be an inexpensive and reliable modality for monitoring nutritive reperfusion status and to obtain prognostic information in patients with AMI.
In previous studies, the incidence of failed reperfusion after thrombolytic therapy has ranged from 15% to 50% (1,2). Of the many diagnostic modalities currently employed for the assessment of reperfusion, coronary angiography has been considered the “gold standard” (3–5). Although several noninvasive markers for reperfusion have been described, the accuracy of these techniques is judged based on their correlation with angiographic findings.
Angiographic assessment is based on the Thrombolysis in Myocardial Infarction (TIMI) flow grade determination in the epicardial infarct related artery (IRA). However, epicardial coronary patency may not mean nutritive cellular perfusion. Additionally, an angiographic “snapshot” of coronary anatomy cannot describe the fluctuations in coronary flow over time that have been reported during the acute phase of myocardial infarction in 35% to 50% of patients (2,6). In contrast, the ST-segment changes observed by electrocardiogram (ECG) monitoring can potentially provide a continuous, noninvasive, real time physiologic marker of nutritive cellular reperfusion.
We hypothesized that electrocardiographically determined reperfusion status might be more predictive of clinical outcomes than angiography. The purpose of this study was to compare the prognostic significance of simultaneous reperfusion assessment by angiographic TIMI flow grade in the IRA versus electrocardiographic ST-segment elevation resolution analysis in patients with acute myocardial infarction (AMI).
All 258 patients who had continuous 12-lead ECG monitoring and protocol acute angiography during the Thrombolytics and Angioplasty in Myocardial Infarction phase 7 (TAMI-7, n = 143) and Global Utilization of Streptokinase tPA for Occluded coronary arteries phase 1 (GUSTO-1, n = 115) trials were included in this study. Inclusion and exclusion criteria for the TAMI-7 and GUSTO-1 trials have been published previously (7,8). All participating patients had provided informed consent for participating in TAMI-7 and GUSTO-1 studies. Inclusion in this study required QRS duration of 120 ms or less. Patients with marked conduction system abnormalities that rendered ECG analysis impossible were excluded from our study.
Data on the following baseline demographic and cardiac risk factor characteristics were provided by the cardiovascular databank at the Duke University Medical Center: time from the onset of symptoms to initiation of thrombolytic therapy, age, height, weight, history of hypertension, diabetes, myocardial infarction, angioplasty, bypass surgery and smoking. Additionally, the peak creatinine phosphokinase (CPK) levels, angiographically determined IRA, left ventricular ejection fraction (LVEF) and the number of major coronary arteries with ≥50% luminal stenosis were included as baseline descriptors.
In accordance with the TAMI-7 and GUSTO-1 protocols, all patients had acute coronary angiography performed (after informed consent) 90–180 min after the start of thrombolytic therapy (7,9). Infarct related artery TIMI flow grade for TAMI-7 patients was determined at the angiographic core laboratory at the University of Michigan Hospital; that for GUSTO-1 patients was determined at the angiographic core laboratory at George Washington University Hospital. For the purposes of our study, patients were divided into three groups: 1) TIMI grade flow 0 & 1, 2) TIMI flow grade 2, and 3) TIMI flow grade 3.
Continuous 12-lead ECG monitoring
All patients in this study had continuous 12-lead ECG monitoring using a portable digital instrument (ST100, Mortara Instrument, Milwaukee, Wisconsin) (2). Monitoring was initiated at median (25th, 75th percentiles) of 3.46 min (−9.33 min, 21.22 min) after the start of thrombolytic therapy and continued for median (25th, 75th percentiles) of 25.17 (12.19, 41.48) hours. All ST-segment monitoring data were analyzed at the Ischemia Monitoring Laboratory, Duke University Medical Center by a single experienced physician, blinded to all clinical and angiographic data. Assessment of each patient’s ST-segment resolution at the time of protocol angiography was performed using the previously published continuous ST-segment resolution analysis method (10). Patients were divided into two groups: 1) those with either less than 50% resolution of the ST-segment elevation from the peak ST-segment elevation in the lead showing the greatest ST-segment elevation, or 150 μV or ≥, and 2) those with 50% or more ST-segment elevation resolution (without reelevation). In order to determine reperfusion status simultaneous with angiography, all analyses were based on continuous ECG data until within 5 min of the initial angiographic contrast injection. The 5 min window was allowed for any possible clock discrepancy. The physician analyzing the ST-segment data was kept blinded to all the continuous ECG data subsequent to the time of initial angiographic contrast injection.
Data for the outcomes were provided by the coordinating center for TAMI-7 and GUSTO-1 trials at Duke University Medical Center. For the purpose of this study, the following parameters were analyzed as outcomes: in-hospital all-cause mortality, the incidence of new congestive heart failure (CHF), and the “combined end point” of mortality or CHF.
Data for baseline and clinical outcome parameters are reported using percentages for categorical variables and 50th (25th, 75th) percentiles for continuous variables. To determine possible univariate correlates of adverse cardiac outcomes and ST-segment/TIMI flow grade category, we applied the chi-square and Fisher’s exact test for categorical variables. A p value of <0.05 was considered statistically significant. The sample sizes and events rates were small in some subgroups; thus, 95% confidence intervals were also given for the event to show how much variability exists. We used exact binomial distribution to find the 95% confidence intervals. A family of comparisons was not used since we were not comparing multiple TIMI flow grades subgroups to one another. We were comparing two ST-segment resolution categories within each TIMI flow grade subgroup. We thought that utilizing 95% confidence intervals was more informative. A logistic regression model was used to determine if ST-segment resolution analysis, rather than TIMI flow grade determination, more accurately predicts the combined end point. Stepwise regression was not used. Multicollinearity was not a factor since TIMI flow was not an independent predictor and stepwise regression was not used.
1. When patients were classified according to TIMI flow grade, outcomes were comparable among all three groups. In patients with TIMI flow grade 0/1, TIMI flow grade 2 and TIMI flow grade 3, the incidence of in-hospital, all-cause mortality was 7.5%, 8.7% and 6.4%, respectively (p = 0.851); the incidence of CHF was 27.5%, 24.6% and 24.8%, respectively (p = 0.894), and incidence of the combined end point was 30.0%, 29.0% and 24.8%, respectively (p = 0.693). When the patients with TIMI flow grades 0/1 and 2 were subdivided according to the ST-segment elevation resolution or persistence, those with the persistence of ST-segment elevation had a striking trend toward worse outcomes in all three parameters, compared with those with the resolution of ST-segment elevation. Almost all the patients (96.3%) with TIMI flow grade 3 had resolution of ST-segment elevation.
2. When patients were classified according to the ST-segment elevation resolution, compared with patients with ST-segment elevation resolution, those with persistent ST-segment elevation had trends toward worse outcomes regarding either the incidence of in-hospital all-cause mortality (6.1% vs. 11.3%, p = 0.174) or the incidence of CHF (23.0% vs. 33.9%, p = 0.085). The incidence of the combined end point was significantly different between the two groups (24.0% vs. 38.7%, p = 0.024).
3. Table 4shows the comparison using the logistic regression models of the ST-segment to IRA TIMI grade flow for the prediction of the combined end point. ST-segment elevation status was an independent predictor of the combined end point (p = 0.024); TIMI grade flow was not (p = 0.693). When both variables were included in a single logistic regression model, ST-segment resolution added additional information TIMI flow grade (p = 0.017). When TIMI grade flow information was added to ST-segment resolution analysis results, no additional prognostic information was gained.
No previous study has compared the relative prognostic significance of reperfusion assessments by angiography versus continuous ECG, using clinical outcomes as the standard. In the past, angiographic findings were used as the “gold standard” to assess the findings of other (noninvasive) techniques. In cases of discordance between angiographic and ECG findings, angiographic findings were considered more reliable. Angiographic epicardial coronary artery patency status may not correspond with the presence or absence of nutritive perfusion at the cellular level. This has been demonstrated in some studies in which discrepancies between outcomes and angiographic evidence of reperfusion were explained as “no-reflow” or “reperfusion injury” phenomena (11,12). Similarly in our study, IRA TIMI flow grade assessment alone was not useful to risk stratify patients, as evidenced by comparable outcomes among patients with all three categories of TIMI flow grade.
The results of our study support the hypothesis that ST-segment resolution as a continuous, physiologic marker of reperfusion predicts outcomes better than a transient, anatomic marker such as angiographic TIMI flow grade in the IRA. Outcomes in our study population did not differ significantly when grouped according to TIMI flow grade although there were trends toward better outcomes in patients with TIMI 3 flow grade compared with those with TIMI grades ≤2 flow. Conversely, when grouped according to ST-segment elevation resolution characteristics, patients whose ST-segment elevation resolved had significantly lower incidence of the combined end point than those whose ST-segment elevation persisted. Similarly, logistic regression analysis indicated ST-segment resolution to be an independent predictor of the combined end point; TIMI grade flow was not an independent predictor and did not add to the information from ST-segment resolution analysis. These results become more significant when one considers that the continuous ECG data for the time after the TIMI flow grade determination were not allowed to be used for patient prognostic stratification. This censoring of available data provided myocardial reperfusion by angiography and ECG, based only on the data obtained until a specific point in time. However, in clinical settings, the continuous ECG monitoring can describe dynamic ST-segment changes even after angiographic determination of coronary anatomy.
Within each TIMI flow grade category (0/1, 2 and 3), patients with ST-segment elevation resolution had trends toward lower incidence of all outcome parameters compared with those with persistent ST-segment elevation. However, due to small numbers in patient subgroups, any concrete conclusions are not possible. These results suggest that when clinical outcomes are used as the standard, TIMI grade 3 flow is a specific, though not a sensitive, marker for successful reperfusion. In our study, angiography failed to distinguish patients with good outcomes among those with TIMI grades <3 flow. However, continuous ECG monitoring was both a specific and a sensitive predictor of clinical outcomes.
It is noteworthy that in our study the in-hospital mortality rate in patients with TIMI flow grade 3 was 6.4%—higher than the 4.4% rate reported in the GUSTO angiographic substudy (9). The reasons for this difference cannot be fully explained but could be due to a smaller sample size or due to the fact that patients from both TAMI-7 and GUSTO-1 trials were included. However, it is important to point out that both mortality rate and the incidence of CHF were always concordant in prognostication of every patient group, emphasizing the validity of outcome data in our population. Similarly, as opposed to our study, GUSTO angiographic study showed TIMI flow grade to correlate with in-hospital mortality. We are unable to fully explain the difference in findings of the two studies. Smaller sample size and inclusion of TAMI-7 patients in our study could, at least in part, explain the difference. It is noteworthy that even in our study, TIMI flow grade 3 was a specific marker for “good outcomes.” However, it was not a sensitive marker, whereas, the resolution of ST-segment was a sensitive as well as specific predictor of “good outcomes.”
TIMI flow grade 2
At the time of the TAMI-7 and GUSTO-1 trials, TIMI flow grade 2 was clinically considered a marker of successful reperfusion. This study considered these patients as a separate group, based on several reports of outcomes in patients with TIMI grade flow 2 intermediate between those with TIMI flow grades 0/1 and 3 and significantly worse than those with TIMI flow grade 3 (13–18). The results of our study suggest that TIMI flow grade 2 is a marker of high risk in some and low risk in others. Although pathophysiology is not considered in this study, in some patients with TIMI flow grade 2, benefits from partial reperfusion may be outweighed by possible risks, such as reperfusion injury. In others, TIMI flow grade 2 would actually be associated with nutritive cellular reperfusion. ST-segment elevation resolution characteristics may help ascertain the true clinical significance of TIMI flow grade 2 in an individual. However, these results regarding TIMI grade 2 flow should be viewed with consideration of the small number of such patients in this population.
Results of our study suggest that “angiographic reperfusion” may not be synonymous with nutritive cellular reperfusion. Angiography does not provide any physiologic markers of cellular reperfusion and only provides brief visualization of epicardial anatomy. Additionally, the invasive and expensive nature of the acute angiography places practical limits on its routine applicability for patients with AMI, especially in the current era of healthcare economics. Obviously, electrocardiographic monitoring is not a substitute for the angiographic data in every patient with AMI. However, in many patients, information obtained through electrocardiographic monitoring can potentially be used to identify patients who need an acute angiogram instead of conservative observation, and in other patients, this information could supplement angiographic data, in acute care and risk stratification.
Although our study was retrospective, it was based on blinded analysis of predefined angiographic, electrocardiographic and clinical variables. The study population was relatively small to address all the issues discussed, characterize other subtle bias and multivariate analysis regarding baseline characteristics. Additionally, collateral blood flow information was unavailable for consideration while making angiography-based decisions. Only 50% level of ST-segment elevation resolution was evaluated in this study. Other degrees of ST-segment resolution may indeed have prognostic values. Because of these limitations, validation of our findings in a larger prospective study will be required.
The findings of our study must be considered preliminary. Assuming validation in a larger patient population, continuous ST-segment monitoring could provide a noninvasive, inexpensive and reliable method for monitoring nutritive reperfusion status in AMI patients. ST-segment monitoring could be utilized for real-time acute patient care and to obtain highly sensitive and specific prognostic information independent of TIMI flow grade in the IRA.
- acute myocardial infarction
- congestive heart failure
- creatinine phosphokinase
- Global Utilization of Streptokinase tPA for Occluded coronary arteries phase 1
- infarct related artery
- left ventricular ejection fraction
- Thrombolytics and Myocardia Infarction phase 7
- Thrombolysis in Myocardial Infarction
- Received August 10, 1998.
- Revision received September 21, 1999.
- Accepted November 15, 1999.
- American College of Cardiology
- Ellis S.G,
- da Silva E.R,
- Heyndrickx G,
- et al.
- Krucoff M.W,
- Croll M.A,
- Pope J.E,
- et al.
- Barbash G.I,
- Roth A,
- Hod H,
- et al.
- Shah P.K,
- Cercek B,
- Lew A.S,
- Ganz W
- Lincoff A.M,
- Topol E.J
- Wall T.C,
- Califf R.M,
- George B.S,
- et al.
- Ito H,
- Tommoka T,
- Sakai N,
- et al.
- Komamura K,
- Kitakaze M,
- Nishida K,
- et al.
- Karagounis L,
- Sorensen S.G,
- Menlove R.L,
- Moreno F,
- Anderson J.L
- Anderson J.L,
- Karagounis L.A,
- Becker L.C,
- et al.
- Vogt A,
- von Essen R,
- Tebbe U,
- Feuerer W,
- Appel K.F,
- Neuhaus K.L
- Kleiman K.S,
- White H.D,
- Ohman E.M,
- et al.