Author + information
- Received August 8, 2005
- Revision received October 5, 2005
- Accepted October 10, 2005
- Published online April 18, 2006.
- Jeptha P. Curtis, MD⁎,
- Edward L. Portnay, MD⁎,
- Yongfei Wang, MS⁎,
- Robert L. McNamara, MD, MHS⁎,
- Jeph Herrin, PhD⁎,
- Elizabeth H. Bradley, PhD†,
- David J. Magid, MD‡,§,
- Martha E. Blaney, PharmD∥,
- John G. Canto, MD, MSPH, FACC¶ and
- Harlan M. Krumholz, MD, SM, FACC⁎,†,#,⁎⁎,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Harlan M. Krumholz, Yale University School of Medicine, 333 Cedar Street, PO Box 208088, New Haven, Connecticut 06520-8088.
Objectives The aim of this study was to determine the use of pre-hospital electrocardiogram (ECG) in patients with ST-segment elevation myocardial infarction (STEMI) undergoing reperfusion therapy, and evaluate the effect of pre-hospital ECG on door-to-reperfusion times.
Background Although national guidelines recommend the use of pre-hospital ECG, there is limited contemporary information about its current use and effectiveness.
Methods Using data from the National Registry of Myocardial Infarction-4, we studied patients with STEMI or left bundle branch block who received acute reperfusion with either fibrinolytic therapy (n = 35,370) or primary percutaneous coronary intervention (PCI) (n = 21,277) within 6 h of admission. We determined the prevalence of pre-hospital ECG use, evaluated the association between pre-hospital ECG and door-to-reperfusion time, and estimated the incremental reduction in time to reperfusion using hierarchical models to adjust for differences in patient and hospital characteristics.
Results A pre-hospital ECG was performed in 4.5% of the fibrinolytic therapy cohort and in 8.0% of the PCI cohort. After adjusting for patient and hospital characteristics, the use of pre-hospital ECG was associated with a significantly shorter geometric mean door-to-drug time: 24.6 min (95% confidence interval [CI]: 23.7 to 25.5) vs. 34.7 min (95% CI: 34.2 to 35.3; p < 0.0001), and a significantly shorter geometric mean door-to-balloon time (94.0 min [95% CI: 91.8 to 96.3] vs. 110.3 min [95% CI: 108.7 to 112.0]; p < 0.0001).
Conclusions The national use of pre-hospital ECG to diagnose and facilitate the treatment of STEMI remains low. When used, however, pre-hospital ECG is associated with a significantly shorter time to reperfusion.
In patients with ST-segment elevation myocardial infarction (STEMI), the rapid administration of reperfusion therapy is associated with improved survival (1,2). Although the American College of Cardiology/American Heart Association guidelines for the management of STEMI recommend door-to-drug times ≤30 min for fibrinolytic therapy and door-to-balloon times ≤90 min for primary percutaneous coronary intervention (PCI) (3), many patients who receive reperfusion therapy do not meet these standards (4–10). The acquisition of a pre-hospital electrocardiogram (ECG) by Emergency Medical Services (EMS) may be an effective method of reducing time to reperfusion (11–16,17).
An analysis from the second National Registry of Myocardial Infarction (NRMI) covering June 1994 to July 1996 demonstrated that the use of a pre-hospital ECG was associated with shorter unadjusted times to reperfusion, a higher proportion of patients receiving reperfusion therapy, and lower in-hospital mortality, although only 5% of STEMI patients received a pre-hospital ECG (18). Since that time, national organizations including the National Heart Attack Alert Program, the American College of Cardiology, and the American Heart Association have promoted the use of the pre-hospital ECG to reduce delays in providing reperfusion therapy (19–23). In addition, increased experience and the development of more reliable methods of performing and communicating results of pre-hospital ECGs may have improved both the use and effectiveness of this strategy (24).
Finally, it is important to understand whether the information from the pre-hospital ECG is used effectively. The longer period of time that the pre-hospital ECG is available before hospital arrival, the greater the opportunity the receiving hospital has to expedite reperfusion therapy by preparing a fibrinolytic agent or alerting cardiac catheterization personnel. In an optimal system, the time that the ECG is available for hospitalization should be inversely related to the time from hospital arrival to reperfusion.
To understand the contemporary use and effectiveness of pre-hospital ECG, we analyzed data from NRMI-4, a large, national, voluntary database that contains detailed information about the pre-hospital ECG and the provision of reperfusion therapy (25,26). We specifically sought to examine changes in the national use of pre-hospital ECG, determine the association of pre-hospital ECG with the time to reperfusion, and determine whether longer availability of a pre-hospital ECG is associated with greater reduction in time to reperfusion.
Study design and sample
This study utilized data on 484,121 patients with acute myocardial infarction (AMI) enrolled in NRMI-4 between January 1, 2000, and December 31, 2002. The criteria for AMI included a diagnosis of AMI according to the International Classification of Diseases, Ninth Revision, Clinical Modification(code 410.X1) and any of the following criteria: total creatine kinase or creatine kinase MB two or more times the upper limit of normal or elevations in alternative cardiac markers; ECG evidence of AMI; nuclear medicine testing, echocardiography, or autopsy evidence of AMI. We restricted our analysis to patients with STEMI or left bundle-branch block (LBBB) who received acute reperfusion therapy with either fibrinolytic therapy or primary PCI <6 h after hospital arrival (Fig. 1).To create this cohort, we excluded patients who had neither ST-segment elevation (2 or more leads) nor LBBB on their first ECG (n = 323,793), patients who were transferred from another institution (n = 40,288), patients whose AMI symptoms began after admission (n = 3,013), patients with unknown symptom onset times or without chest pain (n = 11,301), and patients whose initial ECG was not diagnostic (n = 9,637), missing (n = 1,335), or >6 h after admission (n = 305). Because we were primarily interested in the impact of a pre-hospital ECG obtained by EMS, we excluded patients whose pre-hospital ECG was not performed by EMS (n = 1,105) or was obtained >1 h before arrival (n = 1,874).
From the remaining cohort of 92,575 patients, we excluded those who did not receive primary reperfusion therapy (n = 32,974) or received reperfusion >6 h after presentation (n = 1,849). The final cohort consisted of 56,647 patients who received reperfusion therapy. Of these, we grouped patients into a primary PCI cohort (n = 21,277) and a fibrinolytic therapy cohort (n = 35,370), based on the type of reperfusion therapy first received.
The primary independent variable was the performance of a pre-hospital ECG by EMS. Because a pre-hospital ECG optimally reduces time to reperfusion when there is sufficient time for the results to be communicated to the receiving facility, we grouped patients for whom the pre-hospital ECG was performed <5 min before hospital arrival with patients who did not have a pre-hospital ECG.
The principal outcome of interest was the time between hospital arrival and delivery of reperfusion therapy. In the fibrinolytic therapy cohort, this interval was defined as the time from hospital arrival to the administration of fibrinolytic therapy (door-to-drug). In the PCI therapy cohort, this interval was defined as the time from hospital arrival to first balloon inflation (door-to-balloon). We also assessed the proportion of patients administered fibrinolytic therapy within the American College of Cardiology/American Heart Association guideline-recommended 30-min door-to-drug time and the proportion of patients with prolonged door-to-drug times (>45 min), and those receiving treatment with PCI within the guideline-recommended 90-min door-to-balloon time as well as the proportion with prolonged door-to-balloon times (>120 min). Finally, in the subset of patients in the PCI cohort in whom all time subintervals were documented (n = 21,086), we examined the additional intervals of hospital arrival to arrival in the catheterization laboratory, and arrival at the catheterization laboratory to balloon inflation.
Patient characteristics included as covariates in the analysis included demographic characteristics (gender; age: <65, 65 to 79, ≥80 years; recorded race/ethnicity [white, black, other]; insurance status: Medicare only, Medicare commercial or other, Medicare and Medicaid, commercial, Medicaid only, Veterans Administration, other, self, unknown) and clinical characteristics. Clinical characteristics included medical history (current smoker, chronic renal insufficiency, previous AMI, hypertension, family history of AMI, history of coronary artery disease, hypercholesterolemia, congestive heart failure, previous PCI, previous coronary artery bypass graft surgery, chronic obstructive pulmonary disease, stroke, angina, diabetes), presentation characteristics (chest pain at presentation, systolic blood pressure, pulse, heart failure), and the results of the first ECG after hospital arrival (number of leads with ST-segment elevation, LBBB, AMI location, ST-segment depression, nonspecific ST/T-wave changes, Q-wave). We also included as covariates calendar time (to account for secular trends and staggered reporting periods of hospitals), the reported time between symptom onset and hospital arrival, and arrival time of day and day of week.
Hospital characteristics included as covariates were Census region, a combination of urban/rural location and teaching status, hospital ownership type (government, nonprofit, for-profit), and cardiac facilities (presence of cardiac surgery capability, cardiac catheterization laboratory only, or neither for the fibrinolytic therapy sample). For the door-to-drug analysis, we included the hospital’s annual fibrinolytic therapy volume (0 to 14, 15 to 30, >30 cases) and the percent of all primary reperfusion cases that were done with fibrinolytic therapy rather than PCI (<20%, 20% to 90%, >90%). For the door-to-balloon analysis, we included the hospital’s annual PCI volume (<20, 20 to 40, >40 cases) and percent of all primary reperfusion cases done with PCI rather than drug therapy (<20%, 20% to 90%, >90%). Other hospital characteristics were obtained from the American Hospital Association Annual Survey of Hospitals (27) and the SMG Marketing Group dataset (28). Annual volumes for fibrinolytic therapy and PCI were estimated from the NRMI database.
To determine changes in use of pre-hospital ECG over time, we compared the proportion of patients receiving a pre-hospital ECG in each year of the analysis using the Wilcoxon test for trend. For the primary outcomes of door-to-reperfusion therapy times, we performed separate analyses in the cohort that received fibrinolytic therapy and the cohort that received PCI. We examined how patients who received a pre-hospital ECG differed from those who did not. We compared patient demographics and clinical characteristics as well as hospital characteristics using chi-square tests for categorical variables and ttests for continuous variables.
We evaluated the association between pre-hospital ECG and door-to-reperfusion time and estimated the adjusted door-to-reperfusion time among patients who received and did not receive a pre-hospital ECG using hierarchical regression models that adjusted for patient demographics, clinical characteristics, and hospital characteristics. Hierarchical (also called multi-level) models were necessary because variation in patient outcomes was partly dependent on the hospital to which they were admitted (29). Outcome variables were logarithmically transformed in all analyses to account for the skewness of their distributions (30). Model results were then transformed back to their original units (min) using simulation with 10,000 iterations (31). Because these back transformed effects represent differences in geometric means, adjusted times calculated using these results correspond to adjusted geometric means, as is conventional with log normal outcome measures (30). We specified the intercept and calendar time as random effects, varying across hospitals, to account for hospital-specific differences in outcomes and in change over time.
For the secondary outcomes, we compared the proportion of patients with a pre-hospital ECG who received reperfusion therapy within the guideline-recommended times with the proportion of patients without a pre-hospital ECG who received reperfusion therapy within these timeframes. We then compared the geometric mean of each time interval among patients with and without a pre-hospital ECG in the subcohort of PCI patients who had all intervals available. Finally, we grouped pre-hospital ECG patients by 10-min increments according to the length of time between the performance of the pre-hospital ECG and hospital arrival (≤10, 11 to 20, 21 to 30, 31 to 40, >40 min). We compared the adjusted geometric mean door-to-reperfusion times among these groups using a Wilcoxon test for trend and chi-square test for overall difference. Statistical analyses were performed using SAS version 8.2 (SAS Inc., Cary, North Carolina), and Stata version 8.0 (Stata Corp., College Station, Texas). The institutional review board at the Yale University School of Medicine determined that this protocol was exempt from review because we used existing data that had no patient identifiers.
Use of pre-hospital ECG
Overall, 4.5% (n = 1,599/35,370) of the patients received a pre-hospital ECG in the fibrinolytic cohort, and 8.0% (n = 1,696/21,277) in the PCI cohort. The rate of use of the pre-hospital ECG in both cohorts increased slightly with each consecutive year of the study for patients in both the fibrinolytic cohort (2000: 4.1%, 2001: 4.4%, 2002: 5.1%; p for trend: 0.001) and the PCI cohort (2000: 7.1%, 2001: 7.7%, 2002: 8.8%; p for trend: <0.001).
Patient and hospital characteristics
In both the fibrinolytic and primary PCI cohorts, there were modest but statistically significant differences in sociodemographic and clinical characteristics by whether patients received a pre-hospital ECG (Table 1).The geometric mean time from symptom onset to hospital arrival was shorter for patients who had a pre-hospital ECG than for those who did not receive a pre-hospital ECG in both the fibrinolytic cohort (83 min vs. 95 min) and the PCI cohort (80 min vs. 103 min). Similarly, there were small but statistically significant differences in the characteristics of treating hospitals by whether a pre-hospital ECG was performed (Table 2).
Association of pre-hospital ECG with door-to-reperfusion times
In the fibrinolytic cohort, the geometric mean door-to-drug time was 24.6 min (95% confidence interval [CI]: 23.7 to 25.4) for patients with a pre-hospital ECG and 34.5 min (95% CI: 34.2 to 34.8) for patients without a pre-hospital ECG (p < 0.001) (Table 3).In the PCI cohort, the geometric mean door-to-balloon time was 83.9 min (95% CI: 82.2 to 85.6) for patients with a pre-hospital ECG and 107.7 min (95% CI: 107.0 to 108.3) for patients without a pre-hospital ECG (p < 0.001). After adjusting for hospital-level and patient-level effects with multivariable hierarchical regression, the acquisition of a pre-hospital ECG was associated with a significantly shorter geometric mean door-to-drug time (24.6 min [95% CI: 23.7 to 25.5] vs. 34.7 min [95% CI: 34.2 to 35.3]; p < 0.001) and significantly shorter geometric mean door-to-balloon time (94.0 min [95% CI: 91.8 to 96.3] vs. 110.3 min [95% CI: 108.7 to 112.0]; p < 0.001).
In both the fibrinolytic and primary PCI cohorts, a higher proportion of patients with a pre-hospital ECG had adjusted reperfusion times that fell within the guideline-recommended timeframes compared with patients who did not receive a pre-hospital ECG (Figs. 2Aand 2B). In the fibrinolytic cohort, 60.6% (95% CI: 58.1 to 63.0) of patients with a pre-hospital ECG received fibrinolytic therapy within 30 min of hospital arrival compared with 40.8% (95% CI: 40.3 to 41.3, p < 0.001) of patients without a pre-hospital ECG. In the PCI cohort, 55.2% (95% CI: 52.9 to 57.6) of patients with a pre-hospital ECG received reperfusion therapy within 90 min compared with 33.1% (95% CI: 32.5 to 33.8, p < 0.001) of patients without a pre-hospital ECG.
There were no significant differences between patients in the primary PCI cohort for whom all time intervals were available and those who had missing interval information. Among patients with all time intervals, the adjusted geometric mean door-to-catheterization laboratory-time of those who received a pre-hospital ECG was significantly shorter than that of patients who did not receive a pre-hospital ECG (55.7 min [95% CI: 54.0 to 57.5] vs. 70.5 min [95% CI: 69.2 to 71.8]; p < 0.001). In contrast, the adjusted geometric mean time from arrival at the catheterization laboratory to balloon inflation was comparable among patients with and without a pre-hospital ECG (34.0 min [95% CI: 32.9 to 35.0] vs. 34.6 min [95% CI: 33.8 to 35.3]; p = 0.17).
Longer times between pre-hospital ECG and hospital arrival were not significantly associated with shorter adjusted door-to-drug times (p = 0.63) (Fig. 3).In the PCI cohort, longer times between pre-hospital ECG and hospital arrival were associated with a modest reduction in adjusted door-to-balloon times (p = 0.03). The difference appeared limited to patients in whom the pre-hospital ECG was performed >20 min before hospital arrival.
We found that in spite of guideline recommendations supporting the use of pre-hospital ECG to facilitate the early diagnosis of STEMI, the national use of pre-hospital ECG is low and essentially unchanged from the mid-1990s. When performed, pre-hospital ECGs are associated with both significantly shorter door-to-reperfusion times and a higher proportion of patients receiving reperfusion therapy within guideline-recommended timeframes. Nevertheless, our analysis suggests that the information obtained from pre-hospital ECG is not being used as effectively as possible. These findings support efforts to increase both availability and use of pre-hospital ECG in order to reduce time to reperfusion in AMI patients and highlight the need to develop novel strategies to maximize the potential benefits of pre-hospital ECG.
The main benefit of the pre-hospital ECG lies in its potential to reduce the overall time to administration of reperfusion therapy. Previous studies have demonstrated that patients who have a pre-hospital ECG receive reperfusion therapy faster than patients with no pre-hospital ECG (11,13,15,17,32), and unadjusted analysis of NRMI data from the mid-1990s showed that pre-hospital ECG was associated with a 10-min shorter door-to-drug time and 23-min shorter door-to-balloon time (18). We found that little has changed in the crude door-to-drug and door-to-balloon times associated with the pre-hospital ECG (10-min reduction in door-to-drug time and 24-min reduction in door-to-balloon time). By adjusting for differences in patient and hospital characteristics between patients with and without a pre-hospital ECG, our analyses provide more refined estimates of the actual improvements in time to reperfusion associated with use of pre-hospital ECG compared with the study by Canto et al. Our findings demonstrate that the benefits of pre-hospital ECG have been stable over time. However, they also suggest that there have been no advances in efforts to use the pre-hospital ECG more effectively.
An important difference between the findings of the earlier study and the present one is the interval between symptom onset and hospital arrival. Canto et al. (18) found that patients who received a pre-hospital ECG had a markedly longer symptom onset-to-hospital arrival interval than patients who had no pre-hospital ECG. In contrast, we found that in both cohorts, the geometric mean time between symptom onset and hospital arrival was shorter for patients with pre-hospital ECG. This difference may be due to a number of factors. First, in the present analysis, the interval from symptom onset to hospital arrival was logarithmically transformed to account for the skewed distribution of the data. Second, we were unable to determine whether patients without a pre-hospital ECG arrived via EMS. Accordingly, patients without a pre-hospital ECG in our analysis represent a combination of those who were transported by EMS and those who were self-transported. Studies have demonstrated that patients with acute coronary syndromes transported by EMS present earlier and receive reperfusion therapy faster than patients who are self-transported (33,34). This represents an important unmeasured confounder that may lead to an overestimation of the effect of the pre-hospital ECG on door-to-reperfusion times. Third, changes in practice and improvements in technology may have reduced delays associated with the acquisition of an interpretable ECG. Despite these differences in study methodology, the concordance of the findings strongly suggests that the use of a pre-hospital ECG is associated with shorter door-to-reperfusion times.
Despite national attention to the issue of time to reperfusion (35,36), we found that opportunities to reduce points of delay in the time to reperfusion have been missed. Potential delays in treatment of AMI for patients activating EMS can be divided into four time intervals: 1) symptom onset-to-EMS arrival; 2) EMS arrival-to-hospital arrival; 3) hospital arrival-to-diagnostic ECG; and 4) diagnostic ECG-to-drug/balloon. In theory, the pre-hospital ECG can affect both the third and fourth intervals. By obtaining the ECG before hospital arrival, the third interval (arrival to ECG) is essentially eliminated. In addition, pre-hospital ECG can also markedly reduce the fourth interval (ECG to drug/balloon) because the diagnosis of AMI is made before hospital arrival. Once a pre-hospital ECG is obtained and the results are communicated to the emergency department, there is opportunity to find, prepare, and hang the drug for patients being transported to hospitals that primarily use a fibrinolytic therapy strategy, and the opportunity to notify the interventional cardiologist and prepare the cardiac catheterization laboratory for patients being transported to hospitals that use a primary PCI strategy.
Our findings suggest the potential for even greater reductions in door-to-reperfusion times associated with the routine use of pre-hospital ECG. In the current analysis, we found no significant association between the length of time between the pre-hospital ECG and hospital arrival and subsequent door-to-drug times, and only a modest reduction in door-to-balloon times. These findings suggest that the information from pre-hospital ECG is not being used in the most effective manner. Increased familiarity with pre-hospital ECG, more efficient transmission of pre-hospital ECG results, and adoption of protocols to prepare hospital personnel for the arrival of STEMI patients may all represent opportunities to realize further improvements in door-to-reperfusion times.
Certain aspects of our analysis also merit further consideration. The NRMI database is voluntary and may not include all STEMI patients admitted to participating hospitals (26). In addition, hospitals participating in the NRMI are not representative of all hospitals in the U.S. and include higher proportions of urban, university-affiliated hospitals (25). Finally, we do not have detailed information about the EMS systems providing services to hospitals enrolled in the NRMI. Accordingly, we do not know how the results of the ECG were transmitted or reported to the emergency department. However, numerous studies have shown that reporting or transmitting the ECG to the emergency department is an integral part of most of the pre-hospital ECG programs (11–13,15,16,32,37–40).
Despite nationally accepted quality standards of 30 min for door-to-drug and 90 min for door-to-balloon, many patients do not receive reperfusion therapy within the recommended timeframes (4–9). Increased use of pre-hospital ECG has the potential to reduce time to treatment in that it is associated with substantially shorter door-to-drug and door-to-balloon times. Our findings suggest, however, that pre-hospital ECG is used infrequently, and when used, the information it provides is not being effectively translated into action. Although implementing a pre-hospital ECG strategy requires a significant investment of time, education, and cost, we believe that further efforts to increase the use of pre-hospital ECG, in conjunction with the development of strategies to use this information more effectively could lead to dramatic system-wide reductions in time to reperfusion and ultimately reductions in morbidity and mortality for AMI patients.
This research was supported by the National Heart, Lung, and Blood Institute, grant R01 HL72575-04. Dr. Blaney is employed by Genentech, Inc., South San Francisco, California. Genentech approved the study and provided access to the NRMI database at no charge; however, Genentech did not provide any direct support for the study. The first two authors contributed equally to this work. Dr. Blaney is employed by Genentech, Inc., South San Francisco, California.
- Abbreviations and Acronyms
- acute myocardial infarction
- Emergency Medical Services
- left bundle-branch block
- National Registry of Myocardial Infarction
- percutaneous coronary intervention
- ST-segment elevation myocardial infarction
- Received August 8, 2005.
- Revision received October 5, 2005.
- Accepted October 10, 2005.
- American College of Cardiology Foundation
- Fibrinolytic Therapy Trialists’ (FTT) Collaborative Group
- Antman E.M.,
- Anbe D.T.,
- Armstrong P.W.,
- et al.
- Van de Werf F.,
- Ardissino D.,
- Betriu A.,
- et al.
- Barakat K.,
- Wilkinson P.,
- Deaner A.,
- Fluck D.,
- Ranjadayalan K.,
- Timmis A.
- Mehta R.H.,
- Rathore S.S.,
- Radford M.J.,
- Wang Y.,
- Krumholz H.M.
- Giovas P.,
- Papadoyannis D.,
- Thomakos D.,
- et al.
- Kereiakes D.J.,
- Gibler W.B.,
- Martin L.H.,
- Pieper K.S.,
- Anderson L.C.
- Canto J.G.,
- Rogers W.J.,
- Bowlby L.J.,
- French W.J.,
- Pearce D.J.,
- Weaver W.D.,
- National Registry of Myocardial Infarction 2 Investigators
- Ryan T.J.,
- Antman E.M.,
- Brooks N.H.,
- et al.
- The American Heart Association in collaboration with the International Liaison Committee on Resuscitation
- Lau J.,
- Ioannidis J.,
- Balk E.
- Every N.R.,
- Frederick P.D.,
- Robinson M.,
- Sugarman J.,
- Bowlby L.,
- Barron H.V.
- Rogers W.J.,
- Bowlby L.J.,
- Chandra N.C.,
- et al.
- American Hospital Association
- ↵SMG Marketing Group. Market Overview. Available at: www.smg.com. Accessed March 22, 2006.
- Bryk A.S.,
- Raudenbush S.W.
- Bland J.M.,
- Altman D.G.
- King G.,
- Tomz M.,
- Wittenberg J.
- Canto J.G.,
- Zalenski R.J.,
- Ornato J.P.,
- et al.
- Kudenchuk P.J.,
- Maynard C.,
- Cobb L.A.,
- et al.