Background and Purpose This study describes emergency department (ED) management of stroke in Houston, Tex, in 1992 to identify delays and deficiencies in recognition and management of stroke patients in various hospital subtypes and to quantitate the impact of a rapid response stroke team.
Methods ED logs of eight hospitals were retrospectively screened, and 112 patients with suspected acute stroke onset within 6 hours were identified. EDs were divided into four groups based on hospital size (175 to 979 beds), acuity, number of stroke admissions (50/y to 210/y), and availability of neurological consultations. The intervals from stroke onset to triage, examination by a physician, neurological evaluation, computed tomography (CT) and other tests, vital signs, and treatments were recorded.
Results The average time from stroke to ED arrival was 115 minutes, and times from ED arrival to examination by a physician and CT scan were 28 and 100 minutes, respectively, with little variability among hospital groups except that the public hospital was slower. Neurological examinations were poorly documented in community and public hospitals. The presence of a stroke team shortened the time to examination by a physician and to CT by 13 and 63 minutes, respectively, and increased the number of patients admitted to the intensive care unit. Blood pressure was excessively lowered in 31% of hypertensive patients, and hypotonic dextrose intravenous fluids were given to 69% of all patients.
Conclusions Transport, initial evaluation, and ED care of acute stroke patients are currently slow and often inexpert in all types of hospitals. A stroke team can speed initial ED management.
The Houston Fire Department’s emergency medical system (EMS) responded to approximately 975 stroke-related emergency 911 calls in 1992. These patients were taken to emergency departments (EDs) varying in size, location, level of sophistication, triage numbers, and familiarity with stroke management. This study was undertaken to describe the present state of ED stroke management in representative EDs in our community.
In recent years a greater understanding of the pathophysiology of ischemic stroke and the development of promising new treatments has led to an emphasis on early intervention within the first 6 hours of stroke onset. Animal studies suggest that this 6-hour period is the longest interval of substantial blood flow reduction before maximal and irreversible damage occurs.1 Pathophysiological studies have shown the detrimental effects of hyperglycemia and hyperthermia on the survival of injured brain cells.2 3 The relationship between cerebral blood flow and mean arterial pressure (MAP) is such that acute treatment of high blood pressure can decrease cerebral blood flow, extend infarct size, and decrease the benefits of reperfusion. New treatments to enhance the benefits of early intervention are on the horizon. Neuroprotective drugs that prevent the influx of calcium into the cell can interrupt the chain of chemical interactions that result in cell death. Thrombolytic drugs can dissolve clots that reduce perfusion by obstruction of cerebral vessels. Testing of both types of therapy, alone or in combination, is being carried out in extensive clinical and laboratory research.4 Clinical research protocols require early treatment that is initiated while the patient is still in the ED. Stroke can be caused by hemorrhage as well as by ischemia. Early completion of the computed tomography (CT) scan is needed to identify patients with intracerebral hemorrhage since they may present with a clinical syndrome similar to ischemic infarction but may rapidly deteriorate because of increased intracranial pressure. Therefore, in the first few hours stroke is a dynamic process. Careful attention to changing neurological status, vital signs, and respiratory status can lead to important changes in therapy.
Despite advancements in our understanding of the mechanism of stroke and its treatment, there is little information on existing ED stroke management. The importance of early treatment has been emphasized, and articles have addressed the problem of treatment delays due to lack of public, EMS, and hospital staff awareness.5 Four major points (“the four Ds”) where delay can occur in the ED management of acute myocardial infarction patients were recently described: arrival and triage (door), obtaining an electrocardiogram (data), deciding to treat with thrombolysis (decision), and initiating therapy (drug).6 However, only one recent study on ED management of stroke has been published. This study evaluated the elapsed time between ED arrival and the request for neurological consultation.7 After comparing data from 1986 and 1992, this study concluded that neurological consultations on stroke patients in the ED are delayed more now than in the past.
In this report we describe a study of the existing state of stroke management in the eight hospitals where we are presently conducting a trial of thrombolytic therapy for acute ischemic stroke. The hospitals include a large fully affiliated medical school teaching hospital, large and small private community hospitals with variable teaching roles, and a public urban teaching institution. They range in bed capacity from 175 to 979. We report our evaluation of the differences in stroke care in EDs belonging to hospitals of different size, with varying patient demographics and ED acuity. We also evaluated the impact of the thrombolytic stroke team on the ED management of stroke patients. The purpose of this study was to identify sources of possible delays and errors in the recognition and appropriate management of stroke patients in various ED environments. Although we hope that recognition of such delays and errors might ultimately lead to improved acute stroke treatment, any evaluation of the benefits of acute stroke intervention was beyond the scope and purpose of the present study.
Subjects and Methods
We retrospectively studied the care provided to acute stroke patients in the EDs of eight hospitals in Houston, Tex, during a 12-month period ending December 1992. Patients were identified by screening the ED registration logs of each hospital. Consecutive patients on whom medical records were available with an ED diagnosis of stroke were identified at each ED. The hospitals consisted of 1 medical school hospital, 1 county teaching hospital, 2 community teaching hospitals, and 4 community nonteaching hospitals (Table 1⇓). For comparison purposes, hospitals and their patients were divided into four groups (A, B, C, D) based on hospital size, ED acuity, number of stroke admissions, and availability of neurological consultations. The characteristics of the study hospitals are listed in Table 1⇓. In total, 112 patients were included: 25 patients from the medical school hospital who were screened for entry into the ongoing National Institute of Neurological Disorders and Stroke rTPA (recombinant tissue plasminogen activator) Stroke Trial and consequently received targeted care from the acute stroke team (group A1), 21 patients from the medical school hospital not seen by the stroke team (group A2), 21 patients from the county teaching hospital (group B), 22 patients from the two community teaching hospitals (group C), and 23 patients from the four community nonteaching hospitals (group D).
The study hospitals represent a spectrum of facilities and levels of care within a large urban area of approximately 1.6 million people. Each hospital was served by the Houston Fire Department, whose paramedic services are responsible for transporting most acutely ill patients. All of the EDs were staffed by in-house emergency physicians 24 hours per day. Neurology residents were available for consultation 24 hours per day, 7 days per week at the medical school hospital and on weekdays at the county teaching hospital. All EDs had 2 to 6 shock-trauma beds and 2 to 11 intermediate-care beds. Two of the community nonteaching hospital EDs did not have nonurgent or “holding” beds. The other hospital EDs ranged from 3 to 30 nonurgent beds. CT availability varied (Table 1⇑). A CT scanner was located in the ED of the medical school hospital. At the county teaching hospital, both community teaching hospitals, and two of the community nonteaching hospitals, the CT scanner was immediately adjacent to the ED. The CT scanner was in a separate building at one community nonteaching hospital and was two floors above the ED at the other community nonteaching hospital. The medical school hospital was the only hospital that had a CT technician in-house 24 hours per day, 7 days per week and a laboratory capable of providing immediate results in the ED.
Our stroke team consists of neurologists and nurses who specialize in acute stroke therapy. One physician and one registered nurse from the stroke team were on call 24 hours per day, 7 days per week. When a patient with a probable acute stroke who might qualify for the rTPA Stroke Trial is identified by a paramedic, triage nurse, or ED physician, the stroke team is notified by means of a dedicated beeper system. The physician and nurse on call for the stroke team then meet the patient in the ED. The role of the stroke team is to facilitate emergency stroke care to identify patients who are qualified for the rTPA thrombolytic protocol.
All patients included in this study presented to the ED with signs and symptoms of stroke, were identified as stroke patients by a triage nurse, and were treated for an acute stroke by the ED physicians. To be eligible for this study, all patients had to be at least 18 years old and had to present with onset of stroke signs and symptoms within the previous 6 hours. Patients were excluded if they were intubated en route to the ED, if they presented in a comatose state, or if their stroke was thought to be related to trauma. No patient in this series had a known serious medical illness that would interfere with or complicate treatment of stroke in the ED.
Each patient’s medical record was reviewed to determine the time of onset of symptoms, the time and method (ie, ambulance versus private vehicle) of ED arrival, the time of triage (ED arrival to ED room placement), the time first seen by a physician, and the time of the first documented neurological examination. We also reviewed the vital signs, tests, consultations, and treatments that were ordered and when they were carried out. Treatment of elevated blood pressure was compared with published data relating reduction in MAP to cerebral blood flow.8
All data were collected by one of the principal authors (P.B. or L.G.), and a neurologist (J.C.G.) reviewed each chart to confirm the diagnosis of stroke using established criteria9 and to determine stroke severity using the Scandinavian Stroke Severity Scale.10
Data were entered into a computerized database. Means and standard deviations were calculated. Continuous measures (eg, time intervals) were compared across hospital subgroups by ANOVA. If differences were found, hospital subgroup differences were evaluated by the Student-Newman-Keuls post hoc test. Frequency measures of categorical variables, such as demographics and stroke severity, were compared between hospital groups with the χ2 test of association.
The demographic data reflected the expected age and sex distribution, with little variation among hospital groups. There was a significant variation in race among the hospitals (P<.01) (Table 2⇓), with a larger percentage of white patients at the private community hospitals. More black and Hispanic patients were treated at the medical school and county teaching hospitals. Stroke severity was evenly distributed among all hospital groups.
The stroke diagnoses were divided into ischemic stroke, transient ischemic attack, and intracerebral hemorrhage. Ischemic stroke patients were classified into six categories9 (Table 3⇓). We noted a higher number of ischemic strokes of unknown cause at the county teaching hospital. Eight (7%) of the patients were determined to have “nonstroke” diagnoses at time of discharge from the hospital. The nonstroke diagnoses were subarachnoid hemorrhage (n=2), seizure (n=2), hypoglycemia (n=1), hyperglycemia (n=1), migraine (n=1), and conversion disorder (n=1).
Table 4⇓ lists the mean and standard deviation for each designated interval for the total study population (“all”) and each hospital subgroup. In the total population, the mean time from onset of stroke symptoms to arrival at the ED was 115 minutes; there was no significant difference between patients arriving by ambulance (n=78, 112 minutes) or private vehicle (n=34, 121 minutes). The mean interval from ED arrival to placement into an ED room was 11 minutes. Patients were seen by a physician a mean of 28 minutes after arrival to the ED. Patients arriving by ambulance were examined by the ED physician more rapidly (20 minutes) than those arriving by car (48 minutes) (P<.01). Stroke patients were in the ED for a mean of 123 minutes until they were seen by a neurologist and 100 minutes before a brain CT scan was obtained. The mean time from ED arrival to documentation of vital signs was 7 minutes. Drawing blood took 48 minutes, and obtaining an electrocardiogram took 61 minutes (mean values). Patients stayed in the ED from arrival to disposition for a mean of 324 minutes. Fifty-one percent of patients were admitted to an intensive care unit (ICU). There was no relationship between race or sex and any of the ED intervals we measured.
All hospital EDs proved to be about equal in their efficiency. Teaching hospital EDs were not substantially more or less efficient than other hospital groups. The patients admitted to the county teaching hospital had the longest interval (141 minutes) from onset of stroke symptoms to arrival at the ED, whereas the onset to arrival interval at other hospital types ranged from 101 to 116 minutes. The county teaching hospital also had the longest interval from ED arrival to triage (26 minutes) (P<.01). The county teaching hospital ED also took substantially longer to obtain laboratory tests, other diagnostic tests, and neurological consultation (P<.01).
Neurological examinations were not consistently documented on the initial medical examination at the smaller community nonteaching hospitals compared with all other groups (P<.01). Although there was little difference in the time it took for the patients to be examined by a physician, neurological consultation was fully documented 0% of the time at the community nonteaching hospitals compared with 19% at the county teaching hospital, 32% at the community teaching hospitals, and 87% at the medical school hospital (P<.01).
CT scans were performed while the patient was in the ED in 100% of patients at the medical school hospital, 95% at the county teaching hospital, 86% at the community teaching hospitals, and 74% at the community nonteaching hospitals (P<.01).
At the medical school hospital ED, most intervals studied were shortened when the stroke team was present, including arrival to triage (P=.05) and arrival to CT (P<.01) (Figure⇓). It still took a mean of 50 minutes to complete the neurological examination and CT scan even with the stroke team present. Seventy-six percent of patients seen by the stroke team were admitted to an ICU bed compared with 48% of patients of equal stroke severity who were not seen by the stroke team (P<.01).
Twenty-eight patients (37%) with cerebral infarction had systolic blood pressure (SBP) greater than 170 mm Hg in the ED, and 20 of these patients received antihypertensive therapy while in the ED. The most commonly used drugs were calcium antagonists (13 of 20 [65%]). On review of the ED records (J.C.G.), the use of antihypertensive drugs provoked a precipitous or excessive fall in MAP in 9 of 20 patients (45%) treated for high blood pressure (pretreatment MAP, 137 mm Hg [range, 107 to 150 mm Hg] versus posttreatment MAP, 87 mm Hg [range, 59 to 107 mm Hg]; n=9).
Of the 20 patients with intracerebral hemorrhage, 15 patients (75%) had SBP greater than 170 mm Hg in the ED and all received antihypertensive therapy while in the ED, usually with a calcium antagonist (7 of 15 [47%]) or sodium nitroprusside (4 of 15 [27%]). The use of antihypertensive drugs in 2 of 15 (13%) of these cases caused a precipitous or excessive fall in MAP (pretreatment MAP, 151 and 152 mm Hg versus posttreatment MAP, 93 and 97 mm Hg; n=2).
Eighty-eight patients received intravenous fluids in the ED, and of these, 58 (66%) received solutions that contained dextrose. Sixty-nine percent received hypotonic fluids.
The overall group results confirm that currently EMS, triage, CT scan, neurological consultation, and medical attention for many acute stroke patients are slow. There are many other demands on the resources of paramedics and EDs, but our data indicate certain areas where emergency management of stroke patients may become more efficient. On average, almost 2 hours pass from the time of onset of stroke symptoms until the patient arrives in the ED. Hospital EDs in our study received a relatively small number of code III (life-threatening) patients via EMS. The majority of stroke patients are currently transported as code I (nonurgent) or code II (emergent). This is reflected in the fact that patients arrived at the ED by private car as fast as by ambulance in our study. If acute stroke patients were all given code II or III status, this would probably speed transport and triage.
It takes approximately 30 additional minutes for patient placement into an ED bed and examination by a physician. Interestingly, ED personnel appear to respond more rapidly when a patient arrives by ambulance, an indirect but real advantage of calling 911. Furthermore, in many cases we found that the neurological examination was inadequately documented on the initial record even though stroke was the suspected diagnosis (28 minutes to first documented physical examination but 73 minutes to first documented neurological examination). In many cases a neurological examination may have been carried out before it was documented. In our database, we recorded the first time the neurological findings referable to the stroke appeared in writing on the ED record. In most hospital groups this was within 15 minutes of the first physical examination, but in smaller community hospitals the neurological examination was either not recorded (10 of 23) or substantially delayed. Whether or not complex and potentially dangerous therapies for acute stroke become widely used, a complete neurological examination should occur on the arrival in the ED of any patient suspected of having suffered a stroke. The selection of patients for potentially dangerous new stroke therapies, if they are proven effective, may be problematic in smaller nonteaching community hospitals, where currently we find that neurological evaluations are most inconsistent.
We determined that 25% of our patients had an intracerebral hemorrhage or nonischemic cause for their stroke symptoms, underscoring the importance of CT scanning of all suspected stroke patients. However, on average it took almost 2 hours from ED arrival to obtain a CT scan, and in some hospitals up to 25% of patients did not get a CT scan while they were in the ED. While the availability of in-house CT scan technicians did not per se affect the length of time it took to obtain a CT scan, the ability of a stroke team to produce a 56% reduction in the time needed to obtain a CT scan occurred in a hospital with around-the-clock CT technician availability. It would be more difficult to expedite CT requests in situations in which the technician is on call from home.
Overall, we observed relatively slight variability between our hospital groups regarding the rate at which patients were evaluated. It was disappointing to see that ED management was comparably slow even in medical school and other “teaching” hospitals when the stroke team was not present. One explanation for the greater delay we found in ED management of stroke patients at the county teaching hospital was probably the large number of patients triaged at this facility compared with the other hospitals in our study. In addition, the county teaching hospital had the lowest ratio of neurologists per stroke admission of all the hospitals.
The importance of the availability of neurologists interested in stroke care is underscored by the effect of the stroke team. Comparing patients seen in the ED of our medical school hospital with or without the presence of the stroke team (group A1 versus A2), we found that despite comparable stroke severity and demographics in both groups, the presence of the stroke team speeded all aspects of care. On average, it took 50 minutes to complete the evaluation of a patient and obtain a CT scan with the help of the stroke team. If the time needed for patients to reach the ED is added, the total time to potential treatment is barely within the 3-hour treatment window that might exist for acute stroke therapy. Such therapy would have to significantly reduce mortality or morbidity, however, to justify the added effort of EMS, ED, and stroke team personnel and the higher incidence of ICU admissions that occur as a result of stroke team efforts.
As previously noted by our center and others,8 11 we found that there is still a strong tendency in most EDs to treat elevated blood pressure in patients presenting with acute stroke. Although it is known that MAP will fall spontaneously during the first 24 hours after cerebral infarct,12 71% of cerebral infarct patients with SBP greater than 170 mm Hg were treated in this study, usually with sublingual nifedipine. We have previously shown that a greater than 15% decrease in MAP is associated with a decrease in cerebral blood flow in acute stroke patients,8 yet 45% of treated patients exceeded this degree of MAP reduction. Best management of blood pressure in EDs remains an important target for nurse and physician education.
Best management of intravenous fluids is also an important area of concern. Many stroke patients are dehydrated and need rehydration to maintain optimal cardiac output.13 Most patients in our study received intravenous fluids. In 69% of the cases, the fluids were hypotonic and may have been ineffective for replenishing intravascular volume. Most patients received solutions that contained dextrose. Some evidence suggests that hyperglycemia may augment acute ischemic neuronal damage,2 although the effect of 5% dextrose in intravenous fluids on neuronal survival after stroke is not known.
We believe that our data are representative of stroke care in our community because we included hospitals of variable size, acuity level, number of stroke admissions, and availability of neurological consultation. Furthermore, the subjects of the study represented consecutive admissions to each ED who satisfied our inclusion and exclusion criteria. While we cannot exclude some selection bias in our results, the similarities in our results among the different hospital groups suggest that our data are a reasonable reflection of ED care in our community.
In conclusion, our data show that our current transport, medical evaluation, and ED care of acute stroke patients are frequently slow and inexpert in all types of hospitals. However, most of the problems in ED stroke management can be corrected by faster recognition of stroke symptoms, more prompt patient transport and medical evaluation, the availability of an acute stroke team, and education of ED physicians and nurses.
We would like to acknowledge Howard Rhoades, PhD, for his assistance with statistical analyses and John Marler, MD, for his critical review of the manuscript.
- Received September 13, 1994.
- Revision received November 10, 1994.
- Accepted December 7, 1994.
- Copyright © 1995 by American Heart Association
Kaplan B, Brint S, Tanabe J, Jacewicz M, Wang X, Pulsinelli W. Temporal thresholds for neocortical infarction in rats subjected to reversible focal cerebral ischemia. Stroke. 1991;22:1032-1039.
Yip PK, He YY, Hsu CY, Garg N, Marangos P, Hogan EL. Effect of plasma glucose on infarct size in focal cerebral ischemia-reperfusion. Neurology. 1991;41:899-905.
Ginsberg M, Globus M, Dietrich WD, Busto R. Temperature modulation of ischemic brain injury: a synthesis of recent advances. In: Kogure K, Hossmann K-A, Siesjo BK, eds. Progress in Brain Research. Copenhagen, Denmark: Elsevier Science Publishers; 1993;vol 96.
Grotta JC. The current status of neuronal protective therapy: why have all neuronal protective drugs worked in animals but none so far in stroke patients? Cerebrovasc Dis. 1994;4:115-120.
Feldmann E, Gordon N, Brooks J, Brass L, Fayad P, Sawaya K, Nazareno F, Levine S. Factors associated with early presentation of acute stroke. Stroke. 1993;24:1805-1810.
Malik M, Gomez C, Tulyapronchote R, Malkoff M, Bandlamudi R, Banet G. Delay between emergency room arrival and stroke consultation. J Stroke Cerebrovasc Dis. 1993;3:177-180.
Mohr JP, Caplan LR, Melski JW, Goldstein RJ, Duncan GW, Kistler JP, Pessin MS, Bleich HL. The Harvard cooperative stroke registry: a prospective registry. Neurology. 1978;28:754-762.
Scandinavian Stroke Study Group. Multicenter trial of hemodilution in ischemic stroke: background and study protocol. Stroke. 1985;16:885-890.
Grotta JC, Pettigrew LC, Allen S, Tonnesen A, Yatsu FM, Gray J, Spydell J. Baseline hemodynamic state and response to hemodilution in patients with acute cerebral ischemia. Stroke. 1985;16: 790-795.