Original Contributions

Door-to-ECG time in patients with chest pain presenting to the ED

Deborah B. Diercks MD^a,*, J. Douglas Kirk MD^a, Christopher J. Lindsell PhD^b,c, Charles V. Pollack Jr. MA, MD^d, James W. Hoekstra MD^e,

W. Brian Gibler MD^b, Judd E. Hollander MD^f

^aDivision of Emergency Medicine, University of California, Davis Medical Center, Sacramento, CA 95817, USA

^bDepartment of Emergency Medicine, University of Cincinnati, Cincinnati, OH 45267, USA

^cInstitute for Health Policy and Health Services Research, University of Cincinnati, Cincinnati, OH 45267, USA

^dDepartment of Emergency Medicine, Pennsylvania Hospital, Philadelphia, PA 19104, USA

^eDepartment of Emergency Medicine, Wake Forest University, Winston Salem, NC, USA

^fDepartment of Emergency Medicine, Hospital of the University of Pennsylvania, Philadelphia, PA 19104, USA

Accepted 1 May 2005

Abstract

Objective: To describe time to electrocardiogram (ECG) acquisition, identify factors associated with timely acquisition, and evaluate the influence of time to ECG on adverse clinical outcomes.

Methods: We measured the door-to-ECG time for emergency department patients enrolled in prospective chest pain registry. Clinical outcomes were defined as occurrence of myocardial infarction or death within 30 days of the visit.

Results: Among patients with acute coronary syndrome (ACS), 34% and 40.9% of patients with non- ST-elevation ACS and ST-elevation myocardial infarction , respectively, had an ECG performed within 10 minutes of arrival. A delay in ECG acquisition was only associated with an increase risk of clinical outcomes in patients with STEMI at 30 days (odds ratio, 3.95; 95% confidence interval, 1.06-14.72; P = .04).

Conclusion: Approximately one third of patients with ACS received an ECG within 10 minutes. A prolonged door-to-ECG time was associated with an increased risk of clinical outcomes only in patients with STEMI.

D 2006

Presented at the Society of Academic Emergency Medicine annual meeting, St. Louis MO, May 2002.

T Corresponding author. Tel.: +1 916 734 4052; fax: +1 916 734 7950.

E-mail address: [email protected] (D.B. Diercks).

Introduction

Recent guidelines for the management of patients with acute coronary syndrome (ACS) due to suspected unstable angina or non-ST-elevation myocardial infarction (NSTEMI; collectively, NSTE ACS) recommend that an electrocardiogram (ECG) be obtained within 10 minutes of arrival to the emergency department (ED) in patients with

0735-6757/$ - see front matter D 2006 doi:10.1016/j.ajem.2005.05.016

ongoing chest pain and as soon as possible in all others [1]. This recommendation is based in part on previous reports that show a benefit from rapid ECG acquisition in patients with STEMI [2,3]. Studies in the STEMI population suggest that a prolonged time to ECG acquisition is a significant factor in delayed administration of thrombolytic therapy [3,4].

There are, however, no studies that support a defined door- to-ECG time in the NSTE ACS population. Although there is little evidence that immediate recognition of the NSTE group presents a clear prognostic advantage, it is logical that early identification of these patients can facilitate appropriate and timely management. Furthermore, there are recommenda- tions that support early aggressive treatment in high-risk patients with NSTE ACS, including early percutaneous coronary intervention (PCI) and advanced antiplatelet ther- apy [1]. In addition to the limited knowledge about the value of an early ECG in this group, it is also unclear whether this recommendation can be practically implemented in most institutions. The temporal impact of urgent and timely pharmacological or mechanical interventions on adverse clinical outcomes is still being investigated in this patient population. Likewise, the diagnosis of STEMI or UA/ NSTEMI is largely based on a combination of ECG para- meters and clinical history; therefore, early ECG acquisition is relevant in all patients who present with chest pain consis- tent with a potential cardiac etiology. Therefore, examining our current ECG acquisition practice, identifying factors that influence ECG adherence, and exploring their relationship to adverse clinical outcome are prudent in all populations who present with chest pain.

This study describes our current performance in obtaining an ECG within a 10-minute period and factors associated with early acquisition. It also evaluates the occurrence of adverse clinical outcomes in patients whose ECGs were obtained beyond the recommended 10-minute interval.

Methods

This is a secondary analysis of prospectively collected data. The subjects in this study were enrolled in the Internet Tracking Registry for Acute Coronary Syndromes (i*trACS), a registry of undifferentiated patients with chest pain presenting to 8 geographically dispersed EDs in the United States and Canada. All patients older than 24 years in whom an ECG was obtained in the evaluation of potential anginal symptoms were included. Patients with chest pain associated with cocaine use have been excluded. Three categories of patients were defined based upon final ED diagnoses:

(1) STEMI, (2) UA/NSTEMI, and (3) noncardiac chest pain. All patients within these groups fall under national guidelines that advocate obtaining an ECG within 10 minutes of presentation. Not all of the 8 institutions had established written policies or procedures that were consistent with these guidelines. Institutional review board approval was obtained at all institutions.

Final diagnoses were assigned on a prospective basis using information available to the emergency physician at the time of ED disposition. In this study, STEMI was defined using World Health Organization criteria (ST elevation, classic progression of ST-segment changes, and typical rise and fall of Creatinine kinase-MB in patients with symptoms consistent with a myocardial infarction) and included patients with a new Left bundle-branch block (LBBB) [5]. A final diagnosis of UA was defined as dynamic ECG changes consistent with myocardial ischemia (ie, ST depression N1 mm, transient ST elevation not meeting fibrinolytic cri- teria, or clinical history, in the absence of a positive cardiac injury marker). Non-ST-elevation myocardial infarction was defined as the occurrence of a cardiac marker (creatinine kinase-MB, troponin I, or troponin T) that exceeded the institutional threshold definition for myocardial infarction in patients without ST elevation, but in a patient with clinical ACS. Patients with chest pain without objective evidence of cardiac ischemia or infarction and no other identifiable cause were classified as noncardiac chest pain.

Patient demographics and information on clinical history (including cardiac risk factors), physical examination, and medications administered in the ED (aspirin, nitroglycerin, b-blocker, heparin, and platelet glycoprotein IIb/IIIa inhib- itors) were prospectively collected by trained evaluators at the time of the initial visit or by review of the medical record. history of coronary artery disease (CAD) was based on patient self-report of myocardial infarction, PCI, or coronary artery bypass graft. Confirmation of self-reported CAD was obtained through medical record review when possible. Follow-up was completed on all patients by telephone con- tact, review of the medical record, or social security death registry review. Only clinical outcomes occurring within 30 days were considered.

Patients’ time of arrival and time of initial ECG acquisition were extracted from the database. The exact Mode of transport was not collected in the registry. An Acute cardiac ischemia time-insensitive predictive instrument (ACI-TIPI) score [6] was calculated using extracted data to estimate the risk of coronary ischemia at the time of arrival, although this information was not necessarily available to clinicians in real time. The ACI-TIPI was computed using the most recently available parameter estimates to weight the influence of age, sex, chest pain characteristics, and ECG findings on the probability of cardiac ischemia; the computations were the same as those used in commercially available equipment for computing the ACI-TIPI score. Important adverse clinical outcomes in hospital or during follow-up were defined as either death (all-cause) or recurrent myocardial infarction.

The data were summarized using the median and inter- quartile range for continuous data and frequencies and percentage for categorical data. Patient demographics (age, race, sex, payor) are summarized for the First visit only to prevent bias due to inclusion of multiple visits for some patients. All other variables are summarized for all visits.

	All	STEMI	UA/NSTEMI	Non-ACS	P
	(Npatients = 7488) (Nvisits = 8425)	(npatients = 361) (nvisits = 477)	(npatients = 1035) (nvisits = 1267)	(npatients = 6092) (nvisits = 6681)
Median age (IQR)	53 (43-65)	63 (51-75)	62 (52-73)	51 (42-62)	.085
Race (%) White	3260 (44)	181 (50)	585 (57)	2494 (41)	b.001
African American	3091 (41)	109 (30)	330 (32)	2652 (44)
Other	1137 (15)	71 (20)	120 (12)	946 (16)
Sex (%) Male	3354 (45)	223 (62)	543 (53)	2588 (42)	b.001
Female	4127 (55.)	136 (38)	492 (48)	3499 (57)
Insurer (%) Private	2384 (32)	88 (24)	312 (30)	1984 (33)	b.001
Medicare/Medicaid	2051 (27)	114 (32)	384 (37)	1553 (26)
Self-pay	854 (11)	39 (11)	91 (9)	724 (12)
Other or unknown	2199 (29)	120 (33)	248 (24)	1831 (30)
Medical history (%) Diabetes	1701 (23)	139 (29)	363 (29)	1199 (18)	b.001
Hypercholesterolemia	1977 (26)	114 (31)	474 (43)	1389 (33)	b.001
Angina	1178 (16)	74 (20)	382 (35)	722 (17)	b.001
CAD	2031 (27)	139 (29)	610 (48)	1282 (19)	b.001
CHF	611 (8)	67 (18)	106 (10)	438 (10)	b.001
Family history	3261 (38)	173 (36)	552 (44)	2536 (38)	.039
of CAD
time of visit (%) 8 am to 8 pm (day)	6527 (78)	347 (73)	961 (76)	5219 (78)	.008
8 pm to 8 am (night)	1898 (22.5)	130 (27.3)	306 (24)	1462 (22)
ECG within 10 min (%)	2596 (30.8)	195 (40.9)	428 (34)	1973 (30)	b.001
Median ACI-TIPI score (IQR)	26 (13-37)	38 (24-59)	32 (25-49)	24 (12-32)	b.001
Median time to ECG in min (IQR)	20 (8-52)	14 (5-38)	17 (7-37)	21 (9-56)	.019
Number of clinical outcomes (%)	114 (1)	29 (6)	25 (2)	60 (1)	b.001
Disposition (%) Home	3224 (38)	45 (10)	22 (2)	3157 (47)	b.001
AMA	136 (2)	2 (0)	23 (2)	111 (2)
Admitted	5053 (60)	425 (90)	1221 (96)	3407 (51)
Expired in ED	4 (0)	2 (0)	1 (0)	1 (0)
Median time in ED (5%ile-95%ile)	5.3 (3.7-7.4)	3.8 (2.0-6.0)	5.0 (3.5-7.1)	5.4 (3.8-7.6)	b.001

Differences between subject groups were tested using the Kruskal-Wallis H test for continuous variables and v² tests for categorical variables.

Table 1 Demographics, medical history, and visit information for the 3 patient groups included in this study

Table 2 Adjusted ORs and 95% CIs showing the relationship of patient demographics, medical history, and time of day on whether or not an ECG was acquired within 10 minutes (the constant is not shown in the table)

Independent Reference STEMI UA/NSTEMI Non-ACS

To determine which factors influenced whether or not a patient received an ECG within a 10-minute period, logistic regression was used to compare Patient variables (or groups).

	category	OR	95% CI	P		OR	95% CI	P		OR	95% CI	P
Age		1.00	0.98 -1.03	.771		1.01	1.00 -1.02	.212		1.01	1.00 -1.02	.011
Male	Female	1.08	0.58 -2.01	.820		1.18	0.86 -1.62	.302		1.24	1.05 -1.46	.012
African American	White	0.35	0.18 - 0.67	.002		1.20	0.86 -1.67	.290		0.83	0.70 -1.00	.047
Other race		1.07	0.44 -2.56	.887		1.43	0.79 - 2.57	.236		1.38	1.08 -1.78	.011
Medicare/Medicaid	Private	0.91	0.41-2.04	.823		0.85	0.58 -1.25	.399		0.83	0.67-1.03	.097
Self-pay		2.39	0.88- 6.50	.088		0.71	0.39 -1.30	.270		0.90	0.66 -1.21	.472
Other or unknown		2.34	1.04 - 5.29	.040		1.04	0.67-1.60	.873		1.27	1.02 -1.57	.030
Diabetes	No history	1.19	0.67-2.09	.558		0.96	0.70 -1.32	.801		1.04	0.87-1.24	.667
CAD		1.21	0.68 -2.17	.513		1.31	0.95 -1.80	.105		1.28	1.07-1.53	.008
Hypercholesterolemia		1.62	0.87-2.99	.127		1.17	0.86 -1.60	.321		0.90	0.76 -1.07	.247
Angina		1.27	0.65 -2.48	.476		1.07	0.77-1.47	.697		1.03	0.84 -1.27	.768
CHF		0.98	0.45 -2.12	.954		1.20	0.74 -1.96	.454		1.05	0.80 -1.37	.726
Family history of CAD		0.73	0.41-1.31	.289		0.94	0.68 -1.29	.697		1.02	0.86 -1.20	.842
Night	Day	0.95	0.50 -1.80	.875		1.10	0.77-1.58	.603		1.13	0.94 -1.37	.193
SBP N160 or SBP b90	90 V HR V 160	0.94	0.51-1.73	.835		0.81	0.58-1.12	.203		0.91	0.76-1.09	.287
HR N120	HR V 120	0.70	0.14-3.41	.662		1.17	0.44-3.10	.758		1.22	0.75-2.00	.426

Fig. 1 Histogram showing the time to ECG for each of the patient groups.

Ten minutes was chosen as the standard, based on recent recommendations for the management of UA/NSTEMI, and therefore, time to ECG was entered as a dichotomous variable [1]. For continuous covariates (eg, age, ACI-TIPI), the change in the odds of receiving an ECG within 10 minutes for a unit increase in the variable value is calculated. Logistic regression models were prespecified according to the hypotheses. To determine if institution impacted the results, initial models included this variable. To evaluate the

relationship between time to ECG acquisition and the occurrence of adverse clinical outcomes (All-cause death and recurrent myocardial infarction), logistic regression models were used with time to ECG as a predictor variable. Results have been expressed as odds ratios (ORs) and 95% confidence intervals (CIs) of the ORs. Significance levels are indicated; P b .05 was prospectively set as the limit for statistical significance. Analyses were performed using SPSS for Windows v12.0 and SAS v8.0a.

Table 3 Comparison of patients lost to follow-up and those with a follow-up completed
	MI/LBBB			UA/NSTEMI			CP-NOS
	Lost to	Follow-up		Lost to	Follow-up		Lost to	Follow-up
	follow-up	completed		follow-up	completed		follow-up	completed
Age	63 (51-75)	64 (54-77)		61 (51-72)	61.5 (53-72)		51 (42-63)	50 (41-62)
ACI-TIPI	37.9 (24-59)	34.6 (23-57)		31.6 (25-49)	31.6 (24-52)		24.0 (12-32)	24.8 (13-32)
Female	174 (42)	21 (42)		513 (47)	62 (48)		3504 (60)	247 (52)
Male	237 (58)	29 (58)		576 (53)	68 (52)		2543 (42)	232 (48)
White	202 (49)	29 (58)		601 (55)	64 (49)		2377 (39)	203 (42)
African American	132 (32)	12 (24)		382 (35)	35 (27)		2742 (45)	212 (44)
Other	79 (19)	9 (18)		107 (10)	31 (24)		933 (15)	64 (13)
ECG within 10 min	170 (41)	21 (42)		368 (34)	47 (36)		1817 (30.)	111 (23)
History of diabetes	122 (30)	13 (26)		308 (28)	42 (32)		1086 (18)	84 (18)
History of Hypercholesterolemia	101 (32)	10 (25)		423 (45)	39 (35)		1270 (33)	91 (31)
History of angina	62 (20)	9 (23)		319 (34)	41 (37)		639 (17)	63 (21)
History of CAD	121 (29)	16 (32)		525 (48)	63 (49)		1151 (19)	104 (22)
History of CHF	60 (19)	6 (15)		89 (9)	14 (13)		404 (11)	28 (9)
Family history of heart disease	156 (57)	14 (39)		468 (59)	66 (70)		2290 (57)	190 (57)
Values in bold italics are significantly different at the P b .05 level.

Results

Table 4 Risk of the time to ECG on adverse clinical outcomes (recurrent MI or death in or out of the hospital), corrected for likelihood of ischemia by ACI-TIPI, occurrence of PCI, stent placement or CABG, and treatment

Independent variable OR 95% CI P

The study sample comprised 7887 patients who made a total of 8885 visits to the ED between June 1, 1999, and October 1, 2001. Patients without chest pain were excluded from analyses (399 patients), leaving 7488 patients with 8425 visits. Overall, 575 patients visited the ED twice; 124 patients visited between 3 and 5 times; 7 patients between 5 and 10 times; and 3 patients visited the ED 10 or more times over the 2-year period. Of the 8425 patient visits, 477 (5.7%) had STEMI; 1267 (15%) met the criteria for UA/NSTEMI (1121 UA and 146 NSTEMI); and 6681 (79.3%) were

STEMI
ECG N10 min	3.95	1.06 -14.72	.040
ACI-TIPI	1.02	0.99-1.05	.315
Invasive therapy	1.24	0.32 - 4.77	.753
Noninvasive treatments	1.02	0.10-10.07	.988
History of diabetes	1.18	0.33- 4.24	.801
History of	0.27	0.05-1.33	.106
hypercholesterolemia
History of angina	0.48	0.09-2.53	.390
History of CAD	1.28	0.37- 4.41	.691
History of CHF	2.42	0.52-11.24	.260
Family history	3.13	0.77-12.76	.112
of heart disease
UA/NSTEMI
ECG N10 min	1.13	0.40-3.17	.814
ACI-TIPI	1.03	1.00 -1.06	.023
Invasive therapyT			.996
Noninvasive treatments	1.03	0.13-8.29	.977
History of diabetes	1.04	0.37-2.95	.938
History of	2.18	0.79- 6.05	.133
hypercholesterolemia
History of angina	0.60	0.19-1.85	.371
History of CAD	0.39	0.13-1.16	.090
History of CHF	2.59	0.78-8.58	.119
Family history	1.75	0.59-5.16	.309
of heart disease

classified as noncardiac chest pain. The patient groups are described in Table 1. All predictor variables are significantly different between the 3 patient groups. Patients with the diagnosis of STEMI and UA/NSTEMI were more likely to be white (50.1% and 56.5% respectively), compared with the noncardiac chest pain group (40.9%). There was a statisti- cally significant difference in the time to ECG acquisition between the specific institutions, but the institution was not an independent predictor of clinical outcome after adjusting for time to ECG, and it had no impact on identifying predictors of ECG acquisition. At academic centers, the median time to ECG was 22 minutes, whereas the median time to ECG in the community-based hospital was

12.5 minutes. Complete 30-day follow-up was available in

91.5% of all patients (Table 2).

The median time to ECG acquisition was significantly different among the 3 groups. For patients with STEMI, the

median time to ECG was 14 minutes. Time to ECG was 17 minutes for the UA/NSTEMI group and 21 minutes for	CP-NOS ECG N10 min	1.36	0.68-2.72	.383
those with noncardiac chest pain. Most patients had an ECG	ACI-TIPI	1.01	0.99-1.03	.198
within 30 minutes. There was also a significant difference	Invasive therapy	7.77	3.01-20.10	.000
between groups in the percentage of patients receiving an	Noninvasive treatments	0.56	0.25-1.23	.149
ECG within 10 minutes: 40.9% for the STEMI group, 33.8%	History of diabetes	0.78	0.36-1.69	.531
for UA/NSTEMI, and 29.5% for those with noncardiac chest	History of	0.63	0.29-1.39	.254
pain (Table 1). Most patients had an ECG within 30 minutes (Fig. 1). Determinants associated in logistic regression analysis with whether or not a patient had an ECG performed within the 10-minute time frame are shown in Table 3. For patients with a final diagnosis of UA/NSTEMI, there were no independent predictors of early ECG acquisition. For the group of patients with STEMI, African American race	hypercholesterolemia History of angina 0.67 0.25-1.78 .417 History of CAD 1.10 0.53-2.28 .789 History of CHF 3.68 1.71-7.96 .001 Family history 1.15 0.58-2.28 .686 of heart disease T Because of small cell sizes, the parameters cannot be estimated.

(0.35; 95% CI, 0.18-0.67; P = .002) and unknown insurance status (OR, 2.34; 95% CI, 1.04-5.29; P = .04) were associated with a delayed ECG. For patients with noncardiac chest pain, predictors of delayed ECG were younger age, female sex, African American race, uninsured, and no history of CAD.

The relationship between likelihood of ACS (quantified using the ACI-TIPI score), time to ECG, and the occurrence of an adverse clinical outcome within the 3 patient groups is shown in Table 4. For patients with UA/NSTEMI and non- ACS chest pain, time to ECG was not associated with increased risk of adverse clinical outcome. In the STEMI group, a door-to-ECG time greater than 10 minutes was associated with increased risk of adverse clinical events

(OR, 3.95; 95% CI, 1.06-14.72; P = .04). In the UA/NSTEMI

group, ACI-TIPI score increased the odds of an adverse clinical outcome (OR, 1.03; 95% CI, 1.00-1.06; P = .023). Further modeling shows that the relationship between the risk of adverse events and door-to-ECG time is nonlinear. In the UA/NSTEMI patient population, the proportion of patients with an adverse event was relatively stable up to about

2 hours. After this time, the odds of an adverse event increased significantly. If the ECG was delayed beyond 3 hours, the odds of an adverse event increased 7-fold. (Fig. 2) When evaluating only patients who received an ECG

Fig. 2 Relationship between adverse events and door-to-ECG time for patients with UA or NSTEMI. For the ORs, the reference group is the group with door-to-ECG time less than 5 minutes. Odds ratios are adjusted for disease severity, treatment, and history of CAD or diabetes.

within 2 hours, door-to-ECG time is of borderline statistical significance in predicting adverse events (OR, 3.38; 95%, CI 0.9-12.72; P = .07) among patients with STEMI.

Discussion

The National Heart Attack Alert Program demonstrated that timely acquisition of an ECG is a critical time point (door, data, decision, drug) in the management of patients with STEMI who are undergoing Fibrinolytic therapy [2]. Sagarin et al [3] reported that a delay in ECG acquisition was responsible for 25% of Fibrinolytic treatment delays. Until recently, however, this bdoor-to-dataQ time point has not been emphasized in the patient with chest pain who does not have STEMI. Recent guidelines, however, have recommended that an ECG be obtained in all patients with suspected UA/ NSTEMI and ongoing chest pain within 10 minutes of presentation and as soon as possible in all other patients [1]. This guideline was available at the development of this registry, but in fact, early acquisition has been recognized as an evaluation goal for all patients with chest pain for many years. Recent improvements in treatment options for the patient with UA/NSTEMI based on risk stratification, including ischemic ECG changes, has resulted in an increased urgency in identification of these patients. In addition, the ultimate diagnosis of the undifferentiated patient with chest pain is time dependent, and until an ECG has been performed, the diagnosis of STEMI can not be excluded. This concept supports the early acquisition of an ECG in all patients who present to the ED with possible cardiac chest pain. Therefore, this study examined the current practice in ECG acquisition, identified factors that influenced adherence, and explored their relationship to adverse clinical outcomes in all patients who present to the ED.

The results of our study suggest that current ED practice of obtaining an ECG falls short of the recommended guidelines in all patients with chest pain. We found a median time of

14 minutes for door to data in patients with STEMI. This is similar to the results reported by Jackson et al [7] in 1996, who reported a median door-to-data time of 17 to 25 minutes in patients who received fibrinolytic therapy for STEMI at two institutions. Unlike patients with STEMI, patients with UA/NSTEMI frequently do not have a diagnostic initial ECG. The diagnosis is more often made by cardiac serum markers, serial ECGs, or purely on clinical grounds, typically after an extended evaluation. Despite the difficulty in separat- ing this group of patients from the abundance of patients with non-ACS chest pain at presentation, patients with a final diagnosis of UA/NSTEMI had a median door-to-ECG time of 17 minutes, compared with 21 minutes for patients with noncardiac chest pain.

Prior studies have evaluated potential predictors of a delay in acquisition of an ECG. Lambrew et al [4] reported that sex and mode of transportation were factors affecting the time to ECG in patients with STEMI. Soumerai et al [8] examined the care provided to the elderly between health maintenance organizations or fee-for-service insurance providers and found no difference in time to ECG acquisition. Our findings demonstrate that race and insurance status are predictors of a timely ECG acquisition in patients with STEMI, but that in patients with UA/NSTEMI, there were no independent predictors of an ECG being performed within 10 minutes of presentation.

Rapidly obtaining an ECG is intuitive in any patient with possible coronary ischemia or infarction, and in many institutions, it is part of triage policy and standard ED intake procedure. However, our results suggest that although it may be intuitive, it is not occurring. A broad-based protocol in- corporating a triage ECG could help to correct this. Multiple studies have shown that protocol-driven evaluation of pa- tients with STEMI improves time to ECG and treatment [9-12]. Graff et al [12] derived a triage ECG rule for a rapid 5-minute ECG based on chief complaint, age, and vital signs. The use of this rule decreased the time to ECG from 10 to

6.3 minutes in patients with STEMI, suggesting that a

protocol-driven approach could correct the inadequate door- to-ECG times evidenced here, especially in the UA/NSTEMI population [12].

We found that delayed ECG acquisition was predictive of adverse events only in patients with STEMI, with a 3-fold increase in risk for adverse events in those with an ECG over 10 minutes. This was not noted in patients with UA/NSTEMI where only ACI-TIPI was associated with adverse events. We found that this risk of delayed ECG was apparent predom- inantly above a 2-hour delay in the ECG. It is possible that an atypical presentation of STEMI and UA/NSTEMI resulted in delays in ECG acquisition, subsequently increasing the possibility of adverse outcomes.

This study has several limitations. It is a retrospective analysis of prospectively collected data. Additional data that addressed impediments to ECG acquisition would have been useful and would have provided additional clinical insight into the challenges faced in the ED. For example, we can not account for the physical appearance of patients that may have led to an early ECG. This may have significantly factored into the earlier acquisition in patients with STEMI. It should be noted that we did not account for variability in institutional policies and protocols, although we feel the multicenter for- mat is a strength of this study and improves the generaliz- ability of the results. We also were unable to account for ED overcrowding as a factor that can delay ECG acquisition. Another limitation is the possibility that there may be a discrepancy between the clocks in the ECG machines and ED clocks. This could affect the number of patients having an ECG performed within 10 minutes. We did not account for other comorbidities that may affect ECG acquisition time as well adverse clinical outcome. In addition, we are not able to account for the mode of transportation on door-to-ECG time. Because of the retrospective nature of our study, we could not determine which patients had active pain at the time of presentation or recurrent pain during the visit or those patients with atypical presentation that can lead to delayed ECG acquisition. Including patients with a new LBBB in the STEMI group may have resulted in misclassification of lower-risk patients in this group. This explains why 45 pa- tients were discharged home from the ED with a diagnosis of STEMI. Lastly, because we used ED discharge diagnosis to group patients, we may have had classification error in the patients initially identified as noncardiac chest pain who later were diagnosed with having a myocardial infarction based on subsequent laboratory testing.

In summary, our study shows that some EDs are not meeting the goal for ECG acquisition in patients with UA/ NSTEMI and STEMI and suggests that this may affect adverse clinical outcomes. There appears to be no consistent

predictor of delayed ECG acquisition across all patient categories studied. Recognition of factors that we found to influence time-to-ECG may help others improve perfor- mance. However, improving door-to-ECG time may be dependent upon changing triage criteria and ECG acquisition protocols. This should be followed by further studies to demonstrate if these strategies are effective in changing practice and improving clinical outcome.

References

1. Braunwald E, Antman EM, Beasley JW, et al. ACC/AHA guidelines for the management of patients with unstable angina and Non-ST-segment elevation myocardial infarction. A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee on the Management of Patients With Unstable Angina). J Am Coll Cardiol 2000;36:970 - 1062.
2. Emergency Department. rapid identification and treatment of patients with acute myocardial infarction. National Heart Attack Alert Program Coordinating Committee, 60 Minutes to Treatment Working Group. Ann Emerg Med 1994;23:311 - 29.
3. Sagarin MJ, Cannon CP, Cermignani MS, et al. Delay in thrombolysis administration: causes of extended door-to-drug times and the asymp- tote effect. J Emerg Med 1998;16:557 - 65.
4. Lambrew CT, Bowlby LJ, Rogers WJ, et al. Factors influencing the time to thrombolysis in acute myocardial infarction. Time to Throm- bolysis Substudy of the National Registry of Myocardial Infarction-1. Arch Intern Med 1997;157:2577 - 82.
5. Tunstall-Pedoe H, Kuulasmaa K, Amouyel P, et al. Myocardial infarction and Coronary deaths in the World Health Organization MONICA project. Registration procedures, event rates, and case- fatality rates in 38 populations from 21 countries in four continents. Circulation 1994;90:583 - 612.
6. Selker HP, Beshansky JR, Griffith JL, et al. Use of the acute cardiac ischemia time-insensitive predictive instrument (ACI-TIPI) to assist with triage of patients with chest pain or other symptoms suggestive of acute cardiac ischemia. Ann Intern Med 1998;129:845 - 55.
7. Jackson RE, Anderson W, Peacock IV WF, et al. Effect of a patient’s sex on the timing of thrombolytic therapy. Ann Emerg Med 1996;27:8 - 15.
8. Soumerai SB, McLaughlin TJ, Gurwitz JH, et al. Timeliness and quality of care for elderly patients with acute myocardial infarction under health maintenance organization vs fee-for-service insurance. Arch Intern Med 1999;159:2013 - 20.
9. Bharat V, Tripathy S, Guha AK, et al. Fast track thrombolysis in acute myocardial infarction: a Quality improvement project. Indian Heart J 1998;50:167 - 71.
10. Markel KN, Marion SA. CQI: improving the time to thrombolytic therapy for patients with acute myocardial infarction in the emergency department. J Emerg Med 1996;14:685 - 9.
11. Berglin Blohm M, Nilsson G, Karlsson T, et al. The possibility of influencing components of hospital delay time within emergency departments among patients with ST-elevation in the initial electro- cardiogram. Eur J Emerg Med 1998;5:289 - 96.
12. Graff L, Palmer AC, Lamonica P, et al. Triage of patients for a rapid (5-minute) electrocardiogram: a rule based on presenting chief complaints. Ann Emerg Med 2000;36:554 - 60.