Author + information
- Received August 27, 1996
- Revision received October 31, 1996
- Accepted November 12, 1996
- Published online March 1, 1997.
- John G. Canto, MDA,
- William J. Rogers, MD, FACCA,* (, )
- Laura J. Bowlby, RN, MBAB,
- William J. French, MD, FACCC,
- Douglas J. Pearce, MD, FACCA,
- W.Douglas Weaver, MD, FACCD,
- for the National Registry of Myocardial Infarction 2 InvestigatorsE
- ↵*Dr. William J. Rogers, 334 LHR Building, University of Alabama Medical Center, Birmingham, Alabama 35223.
Objectives. This study sought to examine the management and subsequent outcomes of patients with a prehospital electrocardiogram (ECG) in a large, voluntary registry of myocardial infarction.
Background. The prehospital ECG has been proposed as a means of rapidly identifying patients with acute myocardial infarction who might be eligible for reperfusion therapy.
Methods. The characteristics and outcomes of patients with a prehospital ECG were compared with those without a prehospital ECG in the National Registry of Myocardial Infarction 2 data base. Included in the analysis were those patients who presented to the hospital within 12 h of an acute myocardial infarction. Excluded were patients with an in-hospital infarction, transferred-in referrals and self-transported patients.
Results. Prehospital ECGs were obtained in 3,768 (5%) of 66,995 National Registry of Myocardial Infarction 2 patients meeting study criteria. Median time from myocardial infarction symptom onset until hospital arrival was longer among those having a prehospital ECG (152 vs. 91 min, p < 0.001). However, once in the hospital, the prehospital ECG group experienced a shorter median time to the initiation of either thrombolysis (30 vs. 40 min, p < 0.001) or primary angioplasty (92 vs. 115 min, p < 0.001). The prehospital ECG group was more likely to receive thrombolytic therapy (43% vs. 37%, p < 0.001) and to undergo primary angioplasty (11% vs. 7%, p < 0.001). Also, the prehospital ECG group was more likely to undergo coronary arteriography (55% vs. 40%, p < 0.001), angioplasty (24% vs. 16%, p < 0.001) or bypass surgery (10% vs. 6%, p < 0.001). The in-hospital mortality rate was 8% in patients with a prehospital ECG and 12% in those without a prehospital ECG (p < 0.001). After adjusting for baseline covariates utilizing multiple logistic regression analysis, this mortality difference remained statistically significant (odds ratio 0.83, 95% confidence interval 0.71 to 0.96, p = 0.01).
Conclusions. The prehospital ECG is infrequently utilized for diagnosing myocardial infarction, and among patients with a prehospital ECG, is associated with a longer time from symptom onset to hospital arrival. Despite these shortcomings, the prehospital ECG is a test that may potentially influence the management of patients with acute myocardial infarction through wider, faster in-hospital utilization of reperfusion strategies and greater usage of invasive procedures, factors that may possibly reduce short-term mortality. Efforts to implement the prehospital ECG more widely and more rapidly may be indicated.
(J Am Coll Cardiol 1997;29:498–505)
With the recognition that as many as 1.25 million people will experience an acute myocardial infarction, resulting in almost 500,000 deaths each year in the United States (), there is a paradigm that is clearly shifting toward earlier and more aggressive treatment. By advancing health care delivery through the promotion of government-sponsored national task force recommendations (), community-based educational campaigns and the development of chest pain centers throughout the country, the recent emphasis has been on the early recognition, identification and treatment of acute myocardial infarction.
It is clear that reducing the time to reperfusion therapy in patients with acute myocardial infarction leads to improved survival. The Global Utilization of Streptokinase and t-PA for Occluded Coronary Arteries (GUSTO) angiographic substudy () has shown the critical importance of Thrombolysis in Myocardial Infarction (TIMI) grade 3 flow at 90 min. The Gruppo Italiano per lo Studio della Streptochinasi nell’Infarto Miocardico (GISSI) trial () and other thrombolytic studies ([3, 5–8]), showed a clear inverse relation between time to delivery of thrombolytic therapy and percent myocardial salvage or survival. Thrombolytic therapy within ≤1 h after symptom onset was associated with the greatest benefit. Prehospital thrombolysis has been recently evaluated as a means of decreasing chest pain symptom to treatment times ([8–22]). However, the results from three of the largest prehospital thrombolytic trials ([8–10]) suggest that the critical issue is not where but when therapy is initiated ().
Investigators have begun to examine the utility of the prehospital ECG as a means of rapidly identifying patients with acute myocardial infarction who are eligible for acute reperfusion therapy and thereby reducing in-hospital treatment times. However, to date, limited studies involving only a small number of patients have suggested that the prehospital ECG may decrease in-hospital thrombolytic therapy administration times ([24–28]). Performance of prehospital ECGs is one variable being collected in the ongoing National Registry of Myocardial Infarction 2, a voluntary data base recording presenting characteristics, treatment and in-hospital outcome of patients with acute myocardial infarction at >25% of all acute care hospitals in the United States. The purpose of the present report is to analyze the characteristics and outcomes of patients with a prehospital ECG in this large registry data base.
1.1 Patient population
The National Registry of Myocardial Infarction is a multicenter, voluntary data base designed to collect, analyze and report cross-sectional data on consecutive patients with myocardial infarction enrolled at participating hospitals (). This registry is supported by Genentech, Inc. (South San Francisco, California). Data from each enrolled patient are entered onto a two-page case report form and forwarded to ClinTrials Research, Inc. (Lexington, Kentucky). Before entry into the national data base, all data forms are subjected to internal consistency and data range checking. Participating hospitals are then furnished quarterly tabulations of their individual data as well as parallel quarterly tabulations of state and national results. Strict confidentiality of the individual patient and each respective registry hospital are maintained.
This study reports findings from 1,388 hospitals that enrolled 275,046 patients with acute myocardial infarction into the National Registry of Myocardial Infarction 2 from June 1994 to April 1996. We restricted our analysis to those patients whose first 12-lead ECG was obtained within 12 h of initial myocardial infarction symptoms. Patients with in-hospital infarctions, those transferred in from other hospitals and self (nonambulance)-transported patients were excluded from this analysis.
1.2 Study variables
The study population was classified into two groups: those with a prehospital ECG and those without a prehospital ECG. A prehospital ECGwas defined as any ECG obtained before hospital arrival. The data collection process did not distinguish between prehospital ECGs obtained by paramedics in the field versus those obtained in physician offices.
Patients with a prehospital ECG were compared with those without a prehospital ECG according to the following variables: baseline characteristics, median time interval from myocardial infarction symptom onset to first ECG and hospital presentation; median time intervals from hospital arrival to reperfusion therapy; proportion and modality of acute reperfusion therapy utilized; in-hospital interventions; and clinical events, including mortality. Myocardial infarction symptom onset referred to the onset of cardiac ischemic symptoms related to the acute event and was defined by chest pain or pressure, arm or jaw pain, dyspnea, nausea or vomiting, syncope or cardiac arrest. For patients with stuttering symptoms, onset of infarction was recorded when symptoms became constant in quality or intensity. Also, in a separate analysis, hospital characteristics of participating hospitals contributing at least one patient with a prehospital ECG to the registry were compared with those not enrolling any patients with a prehospital ECG.
1.3 Statistical methods
Differences between the prehospital ECG and no prehospital ECG groups were assessed by the chi-square test for categoric variables, the Wilcoxon rank sum test for time intervals and the ttest for all other continuous variables. Multiple logistic regression analysis was utilized to ascertain variables independently predictive of mortality in the study population. All statistical analyses were performed with SAS 6.06 statistical package programs (SAS Institute). This report is based on information processed by the central data collections center as of April 30, 1996.
Of the 70,763 patients satisfying the study criteria, a prehospital ECG was obtained in only 3,768 (5%). Those patients with a prehospital ECG were slightly younger and were predominantly male, with a lower likelihood of significant cardiac history, such as previous myocardial infarction, angina, heart failure, coronary angioplasty and bypass surgery. They also had a trend toward less heart failure despite a slightly higher frequency of anterior infarction.
2.2 Time intervals: pain onset until ECG, hospital arrival and reperfusion intervention (Fig. 1and Fig. 2).
The median time interval from symptom onset to acquisition of the first ECG was significantly longer in the prehospital ECG group than in the no prehospital ECG group (120 vs. 108 min). In addition, the interval from symptom onset to hospital arrival was longer in the prehospital ECG group (152 vs. 91 min). However, for patients receiving reperfusion therapy, the interval from hospital arrival to treatment was significantly shorter in the prehospital ECG group (Fig. 2): thrombolytic therapy (median time 30 vs. 40 min) and primary angioplasty (median time 92 vs. 115 min). Finally, among those who received either intravenous thrombolytic therapy or primary angioplasty, the overall time from symptom onset to reperfusion therapy was longer in the prehospital ECG group for both thrombolytic therapy (median time 175 vs. 136 min) or primary angioplasty (median time 245 vs. 215 min).
2.3 Acute reperfusion therapy
In the prehospital ECG group, the proportion of patients who received reperfusion therapy was significantly greater (Fig. 3): thrombolytic therapy (43% vs. 37%) and primary angioplasty (11% vs. 7%). Conversely, if a prehospital ECG was not obtained, the patient was more likely not to receive acute reperfusion therapy (56% vs. 46%). In a subgroup analysis, when only patients whose first ECG revealed ST segment elevation or left bundle branch were considered, the overall proportion of patients who received acute reperfusion therapy (thrombolytic therapy or primary angioplasty) was still greater in the prehospital ECG group (71% vs. 65%, p < 0.001).
2.4 Thrombolytic therapy (Table 3).
The physician ordering thrombolytic therapy in the no prehospital ECG group was most often the emergency department physician, whereas in the prehospital ECG group, thrombolytic therapy was ordered with equal frequency between the emergency department physician and the cardiologist. The site of administration of thrombolytic therapy was usually the emergency department in both groups, but patients with a prehospital ECG were somewhat more likely to receive thrombolytic therapy in the intensive care unit or catheterization laboratory. Prehospital thrombolytic therapy was administered in only 0.10% (n = 70) of all study patients.
2.5 In-hospital interventions (Fig. 4).
In the prehospital ECG group, patients were significantly more likely to undergo coronary arteriography (55% vs. 40%), coronary angioplasty (24% vs. 16%) and coronary bypass surgery (10% vs. 6%).
There were no significant differences in recurrent ischemia or recurrent infarction between the two groups, but there were fewer episodes of heart failure and hypotension requiring drug treatment in the prehospital ECG group. In addition, ejection fraction was higher and freedom from adverse events more common in the prehospital ECG group.
In the prehospital ECG group, the unadjusted hospital mortality for the overall population was significantly lower (8% vs. 12%) and for each of the subgroups: intravenous thrombolytic therapy (5% vs. 7%), primary angioplasty (4% vs. 9%) and no reperfusion therapy (12% vs. 16%). To determine whether the prehospital ECG might be an independent predictor associated with lower mortality, multivariate logistic analysis was performed, adjusting for baseline covariates. Variables included in the model are shown in Fig. 5. Prehospital ECG was found to be an independent predictor of lower overall mortality (odds ratio [OR] 0.83, 95% confidence interval [CI] 0.71 to 0.96, p = 0.01).
2.7 Hospital mortality adjusted for symptom onset to first ECG
Because of the disparity in time interval from symptom onset to first ECG (Fig. 1), a secondary analysis was performed assessing mortality as a function of time to first ECG in the prehospital and no prehospital ECG groups. Hospital mortality differed between the two groups only among patients having their first ECG within 3 h of symptom onset and was lower in the prehospital ECG group (OR 0.80, 95% CI 0.65 to 0.99, p = 0.04). The relative benefit of prehospital ECG was even stronger among those receiving their first ECG within 1 h of symptom onset (OR 0.65, 95% CI 0.47 to 0.89, p = 0.007).
2.8 Hospital characteristics (Table 5).
Hospitals in communities with prehospital ECG capability, as defined by the hospital enrolling at least one patient with a prehospital ECG during the study interval, tended to have a larger mean number of staffed beds and emergency department visits per year and were more likely to have coronary arteriography, coronary angioplasty and cardiac surgery capabilities than hospitals not receiving patients with a prehospital ECG. Although prehospital ECGs were obtained in only a minority of patients, >75% of patients (55,290 of 70,763) analyzed in this report presented to hospitals in communities with prehospital ECG capability.
This report, to our knowledge the largest observational study of prehospital ECG to date, shows that the prehospital ECG is infrequently obtained and was acquired in only 5% of 70,763 selected patients with acute myocardial infarction admitted to participating hospitals in the National Registry of Myocardial Infarction 2 data base from 1994 to 1996. Despite its infrequent application, the prehospital ECG was associated with wider usage of both thrombolytic therapy and primary angioplasty; faster in-hospital utilization of such resources; higher usage of invasive procedures, including coronary arteriography and revascularization procedures; and a trend toward lower mortality, even after adjustment for baseline covariates.
3.1 Potential advantages of prehospital ECG
In theory, the prehospital ECG has the potential of rearranging the four major points (the “Four D’s”) where delay can occur in the emergency department (): arrival and triage (Door); obtaining an ECG (Data); decision to use intravenous thrombolytic therapy or alternative reperfusion interventions (Decision); and initiating thrombolytic therapy (Drug). In bypassing “the door” as a barrier to early identification of acute myocardial infarction, the prehospital ECG can be used for prehospital notification that a patient with a myocardial infarction is en route; hospital resources can then be mobilized immediately to rapidly triage the patient for reperfusion therapy on arrival. Early cardiology consultation may be obtained for difficult ECG interpretations, perhaps even before the patient arrives at the emergency department. These scenarios may lead to better patient outcomes.
The prehospital ECG may also be a useful tool to triage patients to hospital facilities better equipped to manage acute myocardial infarction. The surgical community has shown a dramatic improvement in patient outcome after designating hospitals capable of receiving patients for different levels of trauma care ([30–32]). The same critierion may be applied to the patient with an acute myocardial infarction, where rapid and appropriate skilled intervention is of significant importance. Although cardiac patients today are transported to the nearest hospital, the future standard of care may mandate preference (when choice exists) for facilities best equipped to provide state of the art care. The prehospital ECG may potentially be the tool with which this concept is implemented.
3.2 Time to hospital arrival
We anticipated that the prehospital ECG group would receive their first ECG sooner after symptoms onset than the no prehospital ECG group, but this was not the case. Surprisingly, the prehospital ECG group had a significantly longer time from symptoms onset to hospital presentation (152 vs. 91 min). Possible explanations may include the following: 1) a procedural time delay in obtaining the prehospital ECG. Data from three previous studies ([25, 26, 33]) reveal that obtaining the prehospital ECG may add only 5 to 10 min to treatment times. However, although obtained rapidly, the prehospital ECG may trigger other delaying maneuvers in the field, such as starting another intravenous line, drawing up medications or carrying out other orders given by the accepting hospital. In the future, it may still be possible to obtain a prehospital ECG and transport patients faster, with greater paramedic awareness of this potential pitfall. Additionally, the prehospital ECG may have been obtained by staff at private physicians’ offices rather than by paramedics in the field, adding substantially to prehospital delay. 2) The baseline characteristics of the two groups were different, which may affect the speed at which patients are likely to call for help after infarct onset. In this data base, the prehospital ECG group had more patients without a previous cardiac history. Although not proven, such patients may conceivably procrastinate longer before seeking medical attention because they are less likely to recognize the significance of myocardial infarction symptoms. 3) A selection bias may have led to prehospital ECGs being obtained in patients living larger distances from the nearest hospital. 4) A selection bias may have led to greater transport of patients with ECG evidence of an acute myocardial infarction to hospitals recognized for acute myocardial infarction care. Such hospitals may be more distant than the nearest hospital and therefore increase transport times.
Given the unexpected longer presenting time intervals seen in the prehospital ECG group, a subsidiary analysis was performed to further examine the impact of stratifying the data base by onset of symptoms to first ECG time interval. Our results show that obtaining a prehospital ECG within 3 h of symptom onset may favorably influence the management of the patient with an acute myocardial infarction and potentially improve survival. If patients can activate the emergency medical service in an expedient manner, this may allow an opportunity for the prehospital ECG to influence outcome through wider, faster implementation of reperfusion therapy and thereby allow greater salvage of myocardium at risk. In contrast, there may be no additional benefit in obtaining a prehospital ECG if the patient presents much later than 3 h. Presumably, as more myocardium is lost over time, the incremental value of performing a prehospital ECG may also be diminished.
3.3 Time to reperfusion
Although the prehospital ECG group presented much later to the hospital, acute reperfusion therapy was utilized to a greater extent, and median in-hospital interval times to reperfusion therapy were shorter. In-hospital treatment times were less in both the intravenous thrombolytic and primary angioplasty-treated groups, but the disparity was much larger in the latter group. This finding of reduced in-hospital treatment times seems reasonable because many can envision scenarios where the emergency room, aware of the prehospital ECG findings, is ready to receive the patient with a myocardial infarction and administer thrombolytic therapy (so-called “thrombolytic alertness”), or the catheterization laboratory team is prepared to move directly to emergency coronary arteriography and primary angioplasty.
The greater utilization of reperfusion therapy in the prehospital ECG group may be attributed to a higher proportion of patients presenting with ST segment elevation or left bundle branch block on a first ECG as well as a higher proportion of prehospital ECG group patients with an anterior infarction. However, in a separate subgroup analysis that examined only patients with ECG evidence of ST segment elevation or left bundle branch block, the prehospital ECG group still had an overall higher proportion of patients undergoing acute reperfusion therapy. Other plausible explanations may include greater cardiology input on the use of thrombolytic therapy or primary angioplasty in the prehospital ECG group or patients being selectively transported to medical facilities more familiar with indications for reperfusion therapy, once acute myocardial infarction is identified by prehospital ECG testing.
3.4 Limitations of the study
Limitations inherent in the National Registry of Myocardial Infarction data base have been described previously (). There is the potential for nonconsecutive patient enrollment, lack of independent on-site validation of data forms and the fact that registry hospitals may not be representative of all U.S. hospitals. However, the massive size of the registry, its widespread geographical dispersion throughout the United States and its mandatory internal data consistency checks add to its credibility. An additional limitation unique to this study is the inability of the data collection process to distinguish between prehospital ECGs obtained by emergency medical services and those obtained in private physicians’ offices. If many of the prehospital ECGs were obtained in physicians’ offices, this may explain the unexpected longer median times from myocardial infarction symptom onset to acute reperfusion therapy. In addition, the registry did not include details of equipment utilized or the experience and training of personnel obtaining prehospital ECGs. Although such information might have provided useful insights, its absence does not compromise the actual findings of the study. Also, this report represents an observational rather than a randomized analysis. Although a randomized trial is generally better for assessing new diagnostic tools, its use in the prehospital evaluation of outcome may be limited by the Hawthorne effect ([34, 35]). This phenomenon recognizes that study personnel’s behavior may be influenced by the awareness that they are being closely observed in a trial. This limitation of randomized trials is unlikely to affect large registry populations like the National Registry of Myocardial Infarction 2 data base, which reflects “real-world” national practice patterns. Finally, because this registry represented only patients with an actual myocardial infarction, no conclusions can be elucidated in applying this technology to all patients who present with chest pain.
3.5 Clinical implications
These data suggest that the prehospital ECG may be a useful modality for identifying acute myocardial infarction before hospital arrival and is associated with the increased usage of reperfusion interventions; shortened time to implementation of such therapies once the patient reaches the hospital; and the greater utilization of cardiac interventions. The prehospital ECG may represent a test that can influence the management of the patient with an acute infarction and have an impact on patients’ hospital survival. Even with the infrequent utilization of the prehospital ECG at this time, >75% of patients analyzed in this report presented to hospitals in communities with prehospital ECG capability. Although hospitals have an important role in promoting this technology, prehospital emergency medical organizations have the equally important task of properly training more emergency personnel, developing greater awareness of its potential benefits as well as pitfalls and ultimately encouraging its wider usage in the community.
Thus, the prehospital ECG is a modality with great potential for further expansion and may perhaps become a standard of care for guiding treatment of acute myocardial infarction among patients who present to the hospital by ambulance. Despite the favorable outcomes we observed, it should be pointed out that we cannot fully explain nor entirely rule out the possibility that the prehospital ECG itself may be associated with excessive time delay before hospital arrival. Perhaps with faster implementation and wider usage of the prehospital ECG, its potential to facilitate the management of acute myocardial infarction may even be greater. This analysis raises many important questions as to the future direction of prehospital ECG testing in acute myocardial infarction. Subsequent studies are needed to investigate the cost/benefit ratio of universally applying this technology to all patients who present by ambulance with a chief complaint of chest pain.
We are indebted to Stephen L. Webb, Yuan Zhang, PhD, Joseph M. Hilbe, JD, PhD, and Lilly Sanathanan, PhD at ClinTrials Research, Inc. for their statistical expertise.
↵fn1 The National Registry of Myocardial Infarction 2 is supported by Genentech, Inc., South San Francisco, California.
- confidence interval
- electrocardiogram, electrocardiographic
- odds ratio
- Received August 27, 1996.
- Revision received October 31, 1996.
- Accepted November 12, 1996.
- The American College of Cardiology
- ↵National Heart, Lung, and Blood Institute. Morbidity and mortality: chartbook on cardiovascular, lung, and blood diseases. Bethesda (MD): U.S. Department of Health and Human Services, Public Health Service, National Institutes of Health; 1992 Jan.
- The Global Utilization of Streptokinase and t-PA for Occluded Coronary Arteries (GUSTO) Angiographic Investigators
- Tiefenbrunn AJ,
- Sobel BE
- Fibrinolytic Therapy Trialists’ Collaborative Group
- Fine DG, Weiss AT, Welber S, et al. Early prediction of left ventricular function in streptokinase-treated acute myocardial infarction. Cardiology 75:170–6.
- GREAT Group
- GREAT Group
- BEPS Collaborative Group
- Linderer T,
- Schroder R,
- Arntz R,
- et al.
- Bouten MJM,
- Simmons ML,
- Hartman JAM,
- Miltenburg AJM,
- van der Does E,
- Pool J
- the Repair Study Group,
- Bouten MJM,
- Simmons ML,
- Pool J,
- Hartman JAM,
- Miltenburg AJM,
- van der Does E
- Roth A,
- Barbash GI,
- Hod H,
- et al.
- Weiss AT,
- Fine DG,
- Appelbaum D,
- et al.
- Weaver WD
- for the Cincinnati Heart Project Study Group,
- Kereiakes DJ,
- Gibler WB,
- Martin LH,
- Pieper KS,
- Anderson LC
- Valenzuela TD,
- Hamptom DR,
- Keep MH,
- et al.
- Rogers WJ,
- Bowlby LJ,
- Chandra NC,
- et al.
- Roethlisberger FJ,
- Dickson WJ