Time to hospital arrival, use of thrombolytics, and in-hospital outcomes in ischemic stroke

Stroke is a leading cause of disability and death in the United States.¹ The magnitude of stroke can be reduced mainly by effective stroke prevention and treatment of acute stroke. From a treatment perspective, the National Institute of Neurological Diseases and Stroke tissue plasminogen activator trial demonstrated the clinical efficacy of IV alteplase when administered to patients within 3 hours of symptom onset.² Another Phase III, randomized, multicenter trial (Prolyse in Acute Cerebral Thromboembolism II)³ demonstrated the clinical efficacy of intra-arterial pro-urokinase in 180 patients with middle cerebral artery occlusion within 6 hours of symptom onset. A subsequent meta-analysis⁴ involving 18 trials comprising 5,727 patients demonstrated that thrombolytic therapy, administered up to 6 hours after ischemic stroke, significantly reduced the proportion of patients who were dead or dependent at 3 to 6 months. Based on results derived from clinical trials, a substantial effort has been undertaken at both the national and the regional level to increase the proportion of patients who present early after the onset of symptoms.⁵ Education efforts have concentrated on public recognition of symptoms of stroke and importance of early hospital presentation. We performed this regional hospital-based study to 1) provide estimates of patients with ischemic stroke in various strata of time interval between symptom onset and hospital arrival, 2) determine the proportion of patients who receive thrombolysis among those who present in the time window for acute intervention, and 3) determine if the severity of the neurologic deficits and subsequent outcomes are related to time interval between symptom onset and presentation.

Methods.

The Buffalo Metropolitan Area and Erie County Stroke Study was designed to be the first large, regional hospital-based study of temporal trends in stroke incidence rates and outcome within western New York. The Buffalo metropolitan area and Erie County had one of the highest stroke rates in New York State⁶ and therefore represented an ideal site to develop an audit of hospital practices and outcome of stroke. The study involved retrospective identification of all admissions of new or recurrent stroke within the Buffalo metropolitan area and Erie County region between January 1, 2000, and December 31, 2000. The study was performed at 11 hospitals in the study region. Of these, eight were university-affiliated teaching hospitals and three were nonteaching hospitals.⁶ The cases were identified through hospital discharge list and computer-linked record systems to identify patients admitted with stroke. Data on each patient were collected using a standardized data collection form. In brief, study investigators reviewed the medical records of all patients with International Classification of Diseases, 9th revision, Clinical Modification (ICD-9) codes 430 to 438 as primary or secondary discharge diagnoses from the acute-care hospitals in the study region. All the hospitals that serve as possible admission sites for patients with stroke in the Buffalo metropolitan area and Erie County were included in the study. The institutional review committees at each hospital reviewed and approved the protocol prior to initiation of the study. A total of 2,919 cases of stroke were identified using ICD-9 codes. Records were available and reviewed for 2,722 patients (93%).

The criteria that define the various diagnostic categories of stroke, including ischemic stroke subtypes, were adapted from the Classification of Cerebrovascular Diseases III⁷ and from epidemiologic studies of stroke in Rochester, MN.^8–10 Persons with an area of probable infarction on CT scan without any associated clinical symptoms were not included. Strokes were classified as first-ever or recurrent based on evidence of prior clinical stroke in the documentation in the medical records.

The following data were collected for each patient on a standardized questionnaire:

1. Baseline demographics and pertinent data regarding hypertension, diabetes mellitus, cigarette smoking, hyperlipidemia, alcohol use, contraceptive pills, cardiovascular diseases, and previous strokes and any previous use of antiplatelet, anticoagulant, and antihypertensive medications were recorded.⁶

2. Pertinent information was collected regarding the time interval between symptom onset and time to presentation in the emergency room. The time intervals were classified in five categories: 0 to 3, 3 to 6, 6 to 24, and >24 hours and unknown (if time of onset could not be determined).

3. The initial severity of stroke was determined using NIH Stroke Scale (NIHSS) score for all patients with ischemic stroke.¹¹ The NIHSS was estimated by chart abstractors based on the initial recorded neurologic examination. Previously developed instructions for scoring each of the 12 individual NIHSS elements (level of consciousness, response to questions, response to commands, best gaze, visual, facial palsy, motor, ataxia, sensory, best language, dysarthria, and extinction/inattention) were made available to chart abstractors.¹² Missing NIHSS elements were coded as normal.

4. Relevant information was collected regarding the use of IV or intra-arterial thrombolytic treatment in the hospital.

5. At discharge from acute hospitalization, each patient was assigned a grade according to modified Rankin Scale.¹³ This grade was based on the detailed evaluation of rehabilitation medicine physicians or of physical, occupational, and speech therapists participating in the care of the patient. Deaths from all causes during the acute hospital period, inpatient rehabilitation center stays, and chronic supportive care were recorded. Length of hospital stay was determined by the number of days from hospital admission to discharge home or transfer to chronic care or rehabilitation facility.

Results.

The primary diagnosis among the 2,722 patients reviewed included intracerebral hemorrhage (n = 192), subarachnoid hemorrhage (n = 47), TIA (n = 876), and undetermined (n = 17). One thousand five hundred ninety patients of the 2,722 patients presented with ischemic stroke to the study hospitals. The time interval between symptom onset and hospital arrival was 0 to 3 hours in 337 (21%) patients, 3 to 6 hours in 177 (11%) patients, 6 to 24 hours in 301 (19%) patients, >24 hours in 420 (26%) patients, and undetermined in 355 (22%) patients. Table 1 demonstrates the baseline demographic and clinical characteristics of patients according to the strata of time intervals between symptom onset and hospital arrival. No differences were observed in the frequency of various gender and ethnicity categories according to time interval between symptom onset and hospital arrival. IV (n = 23) and intra-arterial (n = 4) thrombolysis was used in 27 (8%) of the 337 patients that presented within 3 hours of symptom onset. For patients presenting between 3 and 6 hours, IV thrombolysis was used in two patients and intra-arterial thrombolysis was used in five patients. IV thrombolysis was used in 1.6% (n = 25) of all 1,590 patients with ischemic stroke admitted to the study hospitals.

Outcome at discharge according to time interval between symptom onset and hospital is presented in table 2. A significantly lower rate of favorable outcome and higher rate of in-hospital mortality was observed among patients with undetermined time interval between symptom onset and hospital arrival (see table 2). There were significant differences observed in the length of intensive care unit stay and hospital stay among different strata of time intervals between symptom onset and hospital arrival as presented in table 2. The mean number of hospital days was greater among patients with undetermined time interval between symptom onset and presentation compared with other time intervals. The mean intensive care unit stay was significantly higher for patients presenting between 0 and 3 hours after symptom onset compared with 3 to 6, 6 to 24, and >24 hours.

In the subset analysis of 1,235 patients with known time interval between symptom onset and hospital presentation, we observed a positive association (p = 0.008) between strata of increasing time interval and proportion of favorable outcomes at discharge (figure 1). There was a nonsignificant relationship between lower in-hospital mortality and strata of increasing time interval (p = 0.09). The intensive care unit stay was related inversely with strata of increasing time interval (p = 0.0001), but no relationship was observed between length of hospital stay and strata of time intervals (p = 0.25). Among the 1,235 patients, initial stroke severity graded according to NIHSS was available in 1,132 patients. The initial NIHSS score was higher with decreasing interval between symptom onset and presentation to the hospital (p < 0.0001) (figure 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. The proportion of patients with favorable outcome defined by modified Rankin Scale score of ≤2 according to time interval between symptom onset and arrival to the hospital.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. The median initial NIH Stroke Scale score with range provided according to time interval between symptom onset and arrival to the hospital.

Discussion.

We observed that approximately 20% of all patients with ischemic stroke presented to the hospital within 3 hours of symptom onset, and only 2% of the 1590 patients received IV or intra-arterial thrombolysis. Recent studies have reported that between 6 and 47% of the patients with ischemic stroke present within 3 hours of symptom onset (table 3).^14–27 Similar or lower rates have been reported in studies conducted outside the United States (see table 3). The proportion of patients that receive thrombolysis is similar to results reported by other community-based studies (see table 3). These studies also highlight the large proportion of patients who present within 3 hours but do not receive IV thrombolysis. The design of our study does not allow us to determine the exact proportion of patients in whom thrombolytics were not administered owing to established exclusion criteria and the reasons that precluded the use of IV thrombolysis in appropriate candidates. We think that the logistical problems relating to rapid clinical or imaging assessment, difficulty in obtaining consent, and physician uncertainty regarding the evidence supporting use of thrombolysis may be responsible for this limited use despite early presentation after symptom onset. A previous study demonstrated that some of these issues can be improved by dedicated efforts, subsequently increasing the proportion of eligible patients and thereby the efficiency of the method.²³

The rate of favorable outcomes at discharge increased with increasing time interval between symptom onset and presentation. This was most likely attributable to higher initial severity of stroke as measured by initial NIHSS score among patients who presented earlier to the hospital after symptom onset. Our results are similar to those derived from International Stroke Trial²⁸ that demonstrated a significant trend for patients with more severe stroke to present earlier than patients with less severe stroke. The investigators observed that the rate of poor outcomes was higher among patients evaluated within 6 hours of symptom onset. This issue may have implications for clinical trials that include patients within 24 hours of symptom onset. Strategies such as randomization stratified by time interval between symptom onset and initiation of treatment need to be considered to avoid imbalances between the treatment groups. In 22% of the patients, time of symptom onset could not be ascertained. These patients were older and had higher rates of in-hospital mortality. This group may represent those patients who presented early but because of elderly age or severity of stroke were not considered candidates for any treatment. Therefore, limited effort was made to ascertain or document the time of symptom onset. The severity of stroke also limits the ability for self-reporting symptom onset. However, we cannot exclude that these patients may represent a heterogeneous group in regard to actual time interval between symptom onset and presentation.

Certain limitations need to be considered prior to interpretation of the results. The first issue was that we used primary and secondary ICD-9 codes to identify patients with stroke admitted to the study hospitals. Previous studies^29,30 suggested that a very small proportion of actual strokes are not identified using both primary and secondary diagnosis ICD-9 codes. The large issue with ICD-9 codes appears to be inclusion of nonstroke patients that was limited by review of all charts and investigations in our study. The second issue was that initial NIHSS was estimated from retrospective review of medical records. Previous studies^12,31 compared prospectively recorded NIHSS with retrospective NIHSS assessment using written history and physical admission notes. Both studies reported a high level of agreement between prospective and retrospective NIHSS score without any systematic bias in retrospective scores. Both studies concluded that for the purposes of retrospective studies of acute stroke outcome, the NIHSS could be abstracted from medical records with a high degree of reliability and validity. Our results are generalizable to community-based settings where hospitals vary in regard to the resources and personnel and there is no mechanism for centralized referral to a specialized center. We acknowledge that the results may not necessarily reflect the outcomes observed in different settings. There may also be an improvement in the rates of thrombolysis over the last 3 years since the study was performed as a result of educational activities conducted for patients, the public, and physicians.

The study highlights the importance of maximizing resources within the hospital to ensure the highest proportion of patients presenting in the time window for intervention receive thrombolytic therapy. Future clinical studies need to consider that patients with more severe neurologic deficits and subsequently worse in-hospital outcomes present early after symptom onset to the hospital.