Early Versus Late Hospital Arrival for Acute Myocardial Infarction in the Western Washington

Thromt@ytic Therapy Trials Charles Maynard, PhD, Ralph Althouse, MD, Michael Olsufka, RN, James L. Ritchie, MD,

Kgthryn B. Davis, PhD, and J. Ward Kennedy, MD

In the 3 Western Washington thrombolytic therapy trials, 54.9% of patjents with acute myocardial infarction arrived at the hospital within 2 hours of symptom onset. These early arrivers were younger and more likely to be hypotensive and in cardiogen- ic shock than were patients arriving later. There were decreases in the time from symptom onset to hospital arrival (p = 0.0002) and in the time from hospital arrival to institution of thrombolytic thera- py (p <O.OOOl) in the 8 hospitals that participated in both the Western Washington intravenous strep- tokinase and tissue plasminogen activator trials from 1983 to 1966. For those patients receiving thrombolysis, early arrival was associated with increased survival (p = 0.031) after adjustment by Cox regression analysis for important clinical pre- dictors of long-term survival. These covariates included pulmonary edema, anterior wall acute myocardial infarction, hypotension and absence of chest pain-at hospital arrival. Reductions in barriers to timely administration’of thrombolytic therapy can be achieved and can result in improved survival.

(AmJCardioj 1989;63:1296-1300)

From the Division of Cardiology, Department of Medicine, School of Medicine, and the Department of Biostatistics, School of Public Health and Community Medicine, University of Washington, Seattle, Wash- ington Manuscript received January 17, 1989; revised manuscript re- ceived February 8, 1989, and accepted March 8.

Address for reprints: J. Ward Kennedy, MD, Division of Cardiolo- gy, RG-22, University of Washington, School of Medicine, Seattle, Washington 98195.

T he importance of early therapy for acute myocar- dial infarction (AMI) has been documented in both laboratory and clinical studies. The interval

between the onset of symptoms and presentation to the emergency room has been shown to -be an important determinant of the efficacy of thrombolytic therapy for AMI.l-3 The objective of this study was to compare ear- ly (<2 hours from symptom onset) and late hospital arrivers (12 hours from symptom onset) with respect to baseline characteristics, 1Cday mortality and long-term survival.

METHODS Patient population: Between July 198 1 and March

1983, 250 patients were enrolled in the Western Wash- ington Randomized Trial of Intracoronary Streptoki- nase in Acute Myocardial Infarction (ICSK),4,5 and be- tween September 1983 and August 1986, 368 patients were participants in the Western Washington Random- ized Trial of Intravenous Streptokinase in Acute Myo- cardial Infarction (IVSK).6 Additionally, 160 patients were enrolled in the Western Washington Tissue Plas- minogen Activator Emergency Room Trial (t-PA) be- tween January 1987 and January 1988. In the first 2 trials, 6 18 patients were randomly assigned to streptoki- nase therapy or to usual coronary care. In the third, nonrandomized trial, 160 patients received intravenous tissue plasminogen activator for the treatment of AMI The effect of thrombolytic therapy for AM1 on short- and long-term survival, measures of left ventricular function and reperfusion has been reported.4-12

This analysis includes 690 (89%) patients enrolled in the 3 trials. Patients already hospitalized at the time of AM1 (61 patients) or those with unknown time to arriv- al (27) were excluded from this analysis. Patients were not eligible for enrollment if time to arrival was > 12 hours in the ICSK trial and >6 hours in the IVSK and t-PA trials. Patients older than 75 years of age were not eligible for inclusion in the trials.

Study variables: The times of symptom onset and hospital arrival were generally available from the emer- gency room flow sheets or ambulance or emergency medical systems reports and were recorded by the study nurse. Symptoms most often referred to chest pain (>90% of patients), although other symptoms such as dyspnea may have brought patients to the hospital. Time to arrival was defined as the time from symptom onset to hospital arrival. Time to therapy was calculated

1296 THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 63

as the time from hospital arrival to the institution of thrombolytic therapy for treatment patients and the in- stitution of other therapies for control patients. In the ICSK trial, all patients underwent cardiac catheteriza- tion as soon as possible after hospital arrival. Conse- quently, time to treatment was calculated for both treatment and control patients. However, IVSK control patients .often did not receive a definite, comparable therapy (for example, heparin) for which time to treat- ment could be determined. Therefore, for many control patients in the intravenous trial, this time was unknown.

Numerous baseline variables were collected and have been defined.4-6 In the t-PA trial additional epide- miologic information not available in the previous 2 tri- als was collected, for example, the mode of arrival to the hospital and patient history of angina, hypertension, cigarette smoking and prior cardiac drug therapy. In the ICSK and t-PA trials, patients were followed 1 year from enrollment, whereas in the IVSK trial, patients were followed at varying intervals with a maximal fol- low-up time >3 years.

Statistical methods: Logistic regression was used to identify the multivariate predictors of early versus late arrival.t3 The effect of time to arrival on survival was analyzed with the log rank statistic14 and Cox re- gression.15

RESULTS The distribution of time to arrival for the 690 pa-

tients included in the analysis is shown in Figure 1. Over half of the patients arrived at the hospital within 2 hours of symptom onset. The percentage of patients ar- riving within 1 hour of symptom onset increased from 12% in the ICSK trial (1981 to 1983) to 16% in the IVSK trial (1983 to 1986) and to 35% in the t-PA trial (1987) (p <O.OOOl), although the upper limit for time to arrival was 12 hours in the ICSK trial and 6 hours in the later trials. However, the percentage of patients ar- riving within 1 hour in the ICSK trial remained the same when patients arriving between 6 and 12 hours were excluded. The percentage of patients arriving within 2 hours of symptom onset was examined by year

TABLE I Characteristics of Early and Late Arrivers

Time from Symptom Onset to Hospital Arrival

<2 hours Variable (n = 379)

Anterior wall AMI 38.0 Inferior wall AMI 62.0 Male gender 83.6 History prior AMI 10.6 Chest pain 91.2 Hemodynamic 84.7

stability Hypotension 13.5 Shock 2.9 Pulmonary edema 5.8 Cardioversion 6.6 Throtibolysis 66.0 Year of enrollment

1981 5.6 1982 17.9 1983 3.4 1984 24.0 1985 17.4 1986 4.2 1987 27.2

Age b-9 55.6 f 10.6 Time to arrival 66.8 f 25.7

(min)

AM = acute myocardial infarction.

22 hours (n = 311)

40.5 59.5 85.5 13.2 91.0 91.6

6.8 0.6 5.1 4.5

58.5

10.3 24.8

5.8 24.1 15.8

6.1 13.2 58.2 f 10.5

207.7 f 104.8

P Value

0.50

0.50 0.44 0.84 0.005

0.004 0.03 0.70 0.24. 0.04. 0.03

0.002

of enrollment, with the result that the percentage of ear- ly arrivers increased with the year of enrollment (p = 0.03, chi-square for trend).

The mean time to arrival for each of the 8 hospitals that participated in the IVSK and t-PA trials is shown in Figure 2. Only hospital C had a significant reduction (p = 0.005), although overall time to arrival was re- duced from 129 f 92 to 94 f 65 minutes (p = 0.0002). It is apparent that in hospitals with shorter arrival times in the IVSK study there was little room for improve- ment, whereas the most improvement was made in hos- pitals with longer arrival times. In these 8 hospitals, 13% of IVSK patients arrived within 1 hour of symp- tom onset.

In Table I, baseline characteristics of patients arriv- ing in <2 hours are compared with those of patients arriving in 12 hours. Early arrivers were significantly

FlQirRE 1. UistribMon of time to hospital anival to symptom onset in the 3 Western Washington thrombolytic therapy tri- als. IC-SK = intracoronary streptoldnase; IV-SK = intra- venous streptokinase; t-PA = tissus plasminogen activator.

FIGURE 2. Mean time to arrival for the 8 hospitals that par- ticipated in both IVSK and t-PA trials. Abbreviations as in Figurel.

THE AMERICAN JOURNAL OF CARDIOLOGY JUNE 1, 1989 1297

HOSPITAL ARRIVAL FOR ACUTE MI

less likely to be hemodynamically stable and, therefore, more likely to be hypotensive or in cardiogenic shock, whereas late arrivers were older and more stable. The mean time to arrival for patients younger than 45 years of age (n = 109) was 101 f 75 minutes, whereas that for patients 65 years of age or older (n = 178) was 138 f 109 minutes (p = 0.002). Early and late arrivers in the t-PA trial did not differ with respect to hist.ory of angina, hypertension, cigarette smoking and prior coro- nary drug therapy. Moreover, the percentage of patients arriving by Medic I (the local emergency medical care system that responds to all 911 calls), ambulance or pri- vate transportation was the same for early and late ar- rivers (p = 0.35).

Stepwise logistic regression was used to identify the predictors of early arrival. From the variables listed in Table I, year of study enrollment (p <O.OOOl), younger age (P = 0.001) and presence of hypotension (p = 0.001) were associated with early arrival. Other vari- ables ‘in Table I such as the presence of cardiogenic shock, which were statistically significant univariate predictors of early arrival, were not identified as such in the multivariate analysis.

Time to treatment did not differ for early and late arrivers who received streptokinase in the IVSK study or those treated with tissue plasminogen activator. It was 99 f 48 minutes for the early and 95 f 56 minutes for the late arrivers in the IVSK study and 52 f 24 minutes for the early and 53 f 23 minutes for the late groups receiving the tissue plasminogen activator. How- ever, for ICSK patients, the time from hospital arrival to cardiac catheterization or initiation of treatment was 153 & 69 and 122 f 8 1 minutes for early and late arriv- ers, respectively (p = O.pO3).

The change in time to treatment for patients treated with thrombolytic therapy was examined for the 8 hos- pitals that participated in both the IVSK and t-PA tri- als. Overall, time to treatment was reduced by almost 40 minutes, from .91 f 44 minutes in the IVSK trial to 52 f 24 minutes in the t-PA study (p <O,OOOl). Figure 3 shows changes in treatment time for the 8 hospitals; the reductions were statistically significant (p <0.05) for all institutions except hospital A (p = 0.10).

FIGURE 3. Mean time to treatment for patients receivtng thrombolytic therapy in the 8 hospiils that participated in both IVSW ad t-PA trials. Abbreviations as in Fii 1.

1298 THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 63

TABLE II Survival for Early and Late Arrivers

Time from Symptom Onset to Hospital Arrival

P <2 hours 12 hours Value

All patients 14 days (%) 93.4 82.9 0.80 1 (%) year 90 a7 0.42 No. of pts 379 311

Thrombolytic therapy 14 days (%) 94.4 93.5 0.72 1 (%) year 92 88 0.42 No. of pts 250 182

I TABLE Ill Variables Entering Cox Regression Analysis for Patients Treated with Thrombolytic Therapy

Infarct location

Model constant

Code

1 = yes 2 = no 1 = anterior 2 = inferior 1 = yes 2 = no 1 = yes 2 = no Minutes

Estimated Coefficient

p Value in Final Model

-1.33 0.0007

-1.03 0.0019

-1.07 0.0061

0.84 0.0357

0.003 0.0309

5.04

1

In Table II, the effect of time to arrival on 1Cday and long-term survival is reported. For all patients, 14- day survival was 93%,for both early and late arrivers (p = 0.80), whereas for patients receiving thrombolytic therapy survival was 94% for both groups (p = 0.72). For patients receiving thrombolytic therapy, l-year sur- vival was 92% for early arrivers and 88% for late arriv- ers (p = 0.42). For all patients; l-year survival was 90% for early and 87% for late-arriving patients (p = 0.42).

To further determine if time to arrival and time to treatment had an effect on survival, Cox regression analysis (using variables in Table I as potential predic- tors) was used. After all significant covariates were al- lowed to enter, the 2 variables were forced into the anal- ysis to determine if they had predictive power. For those receiving thrombolytic therapy, after controlling for the presence of pulmonary edema, anterior wall AMI, pres- ence of hypotension and absence of chest pain on hospi- tal arrival, increased time to arrival was associated with decreased survival (p = 0.031) (Table III). This finding lost its significance when control patients were added to the analysis (p = 0.31), nor was there a statistically significant association between time to treatment and survival (p = 0.43).

DISCUSSION Of note, the proportion of patients arriving at the

hospital within 1 hour of symptom onset increased dra- matically since 198 1. This increase was most evident in the period from late 1983 to early 1988, as evaluated in the 8 hospitals that participated in both intravenous trials.

There are several possible explanations for this find- ing. It is likely that the general public has become more aware of the need for prompt action in response to symptoms suggestive of AMI. During this time period, there was a concerted media campaign directed by the American Heart Association of Washington to educate the public about recognition of symptoms of AM1 and the need for prompt hospitalization. Second, although we have controlled for institution, it is possible that changes associated with the hospitals may have resulted in decreased time to arrival. For example, there may have been changes in the emergency medical systems at these institutions that resulted in faster arrival times. Third, it is possible that changes in patient mix in the areas served by these institutions could influence time to hospital arrival. For example, if there were an increase in patients who were hypotensive, there could be a cor- responding decrease in time to arrival. However, pa- tients in the 3 trials were very similar with respect to baseline characteristics such as age, sex, hypotension and cardiogenic shock.

The finding that older patients with AM1 presented later whereas patients with more severe symptoms ar- rived at the hospital earlier has been reported by the Multicenter Investigation for the Limitation of Infarct Size.16 Other investigators have also noted trends in the relation between age and late hospital arrival for AMI.17J8

Studies that have analyzed the delays of predomi- nantly white patients in seeking medical care for myo- cardial infarction19-** were compared by Cooper et a123 with the mostly black population of Cook County Hos- pital. The comparison showed a much greater delay in requesting assistance among black patients. Because race was not reported in the Western Washington trials, it was not possible to examine racial differences in arriv- al times.

It is also possible that history of angina pectoris could be associated with time to’ arrival because patients may have difficulty deciding whether chest pain is angi- na pectoris or AMI. Information about the history of angina pectoris was not collected in the ICSK and IVSK trials. This information was available in the t-PA trial, in which history of angina was not related to time to arrival, even after controlling for age. More recently, Ho et a124 have reported on why patients with chest pain delay in calling 9 11.

It is also important to note the change in delays be- tween hospital arrival and initiation of therapy. Therapy in the ICSK trial required additional time for prepara- tion of the cardiac catheterization laboratory, and took about 30 minutes longer to institute than did therapy in the IVSK trial. Nevertheless, it still took >90 minutes to begin treatment with IVSK once the patient arrived at the hospital. This time varied considerably by hospi- tal, with a range of 80 to 144 minutes. A portion of this time was required to obtain informed consent for entry into the trial. When the protocol for the nonrandomized t-PA trial was written, conscious efforts were made to reduce time to treatment to <l hour. Emergency room and coronary care personnel attended in-service training

sessions conducted by study nurses, The average time to treatment was reduced to 52 minutes (range 43 to 72). There are several possible explanations for this change, although it is most likely that the study protocol re- quirements (including the nonrandom nature of the trial and in-service training sessions) were responsible for the reduction.

However, patients enrolled in these 3 trials may be different from other patients who have AMI. First, these patients survived long enough to reach the hospi- tal, whereas a significant percentage of patients with AM1 die before reaching the hospital. Second, these pa- tients arrived in $12 (ICSK) or <6 (IVSK, t-PA) hours of symptom onset and patients with longer arrival times were nat included. Age differences with respect to time to arrival may have been even more accentuated had patients older than 75 years been included. Also, Western Washington patients were distinguished by the fact that they gave informed consent to participate in the trials, 2 of which required random allocation to therapy. Consequently, the applicability of these find- ings may be further limited. Nevertheless, the associa- tion between age and hospital arrival, and symptom se- verity and hospital arrival is supported by other studies.

Although time to arrival was not, by itself, an impor- tant predictor of survival, it did become so after adjust- ing for other known predictors of survival. Furthermore, faster arrival times were associated with increased sur- vival for patients receiving thrombolysis but not for those treated conventionally. The findings of this analy- sis suggest that barriers to rapid hospital arrival and treatment can be reduced and that the payoff is longer life. The Western Washington trials were not designed to identify and change these barriers, though it is appar- ent that their effects were attenuated. Current trials are seeking to further reduce barriers to rapid treatment by starting thrombolysis in emergency medical vehicles.25 The reduction of barriers to early arrival requires im- proved efforts with respect to community education as well as a responsive emergency medical care system.

REFERENCES 1. The GISSI Study Group. Effectiveness of intravenous thrombolytic therapy in acute myocardial infarction. Lancer 1986;1:397-401. 2. The ISAM Study Group. A prospective trial of intravenous streptohinase in acute myocardial infarction (ISAM): mortality, morbidity, and infarct size at 21 days. N Engl J Med 1986;314:146.5-1471. 3. The ISIS-2 Collaborative Group. Randomised trial of intravenous streptoki- nase, oral aspirin, both, or neither among 17187 cases of suspected myocardial infarction: ISIS-2. Lmcet 1988;2:349-360. 4. Kennedy JW, Ritchie JL, Davis KB, Fritz JK. Western Washington Random- ized Trial of Intracoronary Streptokinase in Acute Myocardial Infarction. N Engl J Med 1983;309:1477-1482. 5. Kennedy JW, Ritchie JL, Davis KB, Stadius ML, Maynard C, Frit? JK. The Western Washington Randomized Trial of Intracoronary Streptokinase in Acute Myocardial Infarction: a 12 month followup report. N Engl J Med 1985;312: 1073-1078. 6. Kennedy JW, Martin GV, Davis KB, Maynard C, Stadius ML, Sheehan FH, Ritchie JL. The Western Washington Intravenous Streptokinase in Acute Myo- cardial Infarction Randomized Trial. Circulation. 1988;77:345-352. 7. Althouse R, Maynard C, Olsutlra M, Ritchie JL, Kennedy JW. The Western Washington Tissue Plasminogen Activator Emergency Room Study (abstr). JACC 1989;13:94A. 6. Ritchie JL, Davis KB, Williams DL, Caldwell J, Kennedy JW. Global and regional left ventricular function and tomographic radionuclide perfusion. The

THE AMERICAN JOURNAL OF CARDIOLOGY JUNE 1, 1989 1299

HOSPITAL ARRIVAL FOR ACUTE MI

Western Washington Intracoronary in Myocardial Infarction Trial. Circulation 1984;70:867-875. 9. Stadius ML, Maynard C, Fritz JK, Davis K, Ritchie JL, Sheehan F, Kennedy JW. Coronary anatomy and left ventricular function in the first 12 hours of acute myocardial infarction: the Western Washington Randomized Intracoronary Streptokinase Trial. Circulation 1985;72:292-301. 10. Stadius ML, Davis K, Maynard C, Ritchie JL, Kennedy JW. Risk stratifica- tion for 1 year survival based on characteristics identified in the early hours of acute myocardial infarction. The Western Washington Intracoronary Streptoki- nase Trial. Circulation 1986;74:703-711. 11. Martin GV, Stadius ML, Davis KB, Ritchie JL, Sheehan F, Maynard C, Kennedy JW. The Western Washington Streptokidase Trial: effects of intrave- nous streptokinase on vessel patency and LV function. Circulation 1988;78:258- 266. 12. Ritchie JL, Cerqueira M, Davis K, Maynard C, Kennedy JW. Western Washington Intravenous Streptokinase in Myocardial Infarction Trial: radionu- elide ventricular function and infarct size. JACC 1988;11:689-697. 13. Kleinbaum DG, Kupper LL, Morgenstern H. Epidemiologic Research: Prin- ciples and Quantitative Methods. Belmont, California: Lifetime Learning Publi- cations, 1982:434-446. 14. Kalbfleisch JD, Prentice RL. The Statistical Analysis of Failure Time Data. New York: John Wiley, 198O:l O-l 9,70-86. 15. Cox DR. Regression models and life tables. JR Stat Sot [B] 1972;37:189- 200. 16. Turi ZG, Stone PH, Muller JE, Parker C, Rude RE, Raabe DS, Jaffe AS, Robertson TL, Braunwald E, the MILIS Study Group. Implications for acute intervention related to time of hospital arrival for acute myccardial infarction. Am

J Cardioi 1986;58:203-209. 17. Hackett TP, Cassem NJ. Factors contributing to delay in responding to the signs and symptoms of acute myocardial infarction. Am J Cardioll969;24:651- 658. 18. Moss AJ, Wynar B, Goldstein S. Delay in hospitalization during the acute coronary period. Am J Cardiol 1969:24:659-665. 19. Pressley JC, Wilson HB, Severance HW, Raney MP, McKinnis RA, Smith MW, Hindman MC, Wagner G. Basic emergency medical care of patients with acute myocardial infarction: initial pm-hospital characteristics and in-hospital complications. JACC 1984;4:487-492. 20. Zmyslinski RW, Lackland DT, Keil JT, Higgins JE. Increased fatality and difficult diagnosis of in-hospital acute myocardial infarction: comparison to lower mortality and more easily recognized prehospital infarction. Am Heart J 1981;101:586-592. 21. Schroeder JS; Lang IH, Hu M. The prehospital course of patients with chest pain. Am J Med 1978;64:742-748. 22. Feinleib M, Davidson MJ. Coronary heart disease mortality: a community perspective. JAMA 1972;222:1129-1134. 23. Cooper RS, Simmons B, Castaner A, Prasad R, Franklin C, Felinz J. Survival rates and prehospital delay during myocardial infarction among black persons. Am J Cardiol 1986;57:208-211. 24. Ho MT, Eisenberg MS, Litwin P, Schaeffer S, Damon S, Weaver WD, Hallstrom A, Larsen MP. Reasons chest pain patients delay or do not call 911 (abstr). Circulation 1988;78:11-87. 25. Weaver WD, Martin JS, Litwin P, Eisenberg MS, Ho MT, Kudenchuk PJ, Hallstrom AP. Prehospital thrombolytic therapy-preliminary report of feasibil- ity (abstr). Circulation 1988;78:II-111.

1300 THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 63