- Email: [email protected]
Clinical Outcomes among Stroke Patients Receiving Tissue Plasminogen Activator Therapy Beyond the 3-hour Time Window
Clinical Outcomes among Stroke Patients Receiving Tissue Plasminogen Activator Therapy Beyond the 3-hour Time Window Emily C. O’Brien, MSPH, Kathryn M. Rose, PhD, Mehul D. Patel, MPH, Carol V. Murphy, RN, MPH, and Wayne D. Rosamond, PhD
Background: Tissue plasminogen activator therapy (t-PA) is associated with improved neurologic outcomes and reduced disability from ischemic stroke. The current guidelines stipulate that patients receive t-PA within 3 hours of symptom onset. However, actual practice patterns vary, and little is known about patient outcomes when t-PA is received outside of the recommended time window. Methods: We examined mean length of hospital stay, t-PA–related complications, and in-hospital death by time of t-PA administration in North Carolina Stroke Care Collaborative (NCSCC) patients. The NCSCC includes 53 hospitals that enroll patients presenting with stroke-like symptoms. Of 40,907 patients enrolled between January 2005 and February 2010, 1070 (2.6%) received t-PA. Of these, 88.2% received t-PA within 3 hours of symptom onset (‘‘early’’) and 30.3% received t-PA between 3 and 6 hours after symptom onset (‘‘late’’). Results: Unadjusted mean length of stay (days) was longer among early patients (5.0 days; 95% confidence interval [CI], 4.7-5.3) than late patients (3.6 days; 95% CI, 3.1-4.2). t-PA–related complications were similar among early (7.0%; 55/781) and late patients (6.7%; 7/102; P 5.89). The proportion of in-hospital deaths was similar among late (10.5%) and early patients (12.0%). We used multivariable logistic regression to estimate odds ratios (ORs) and 95% CIs for the associations between late t-PA status and patient outcomes. Conclusions: In models controlling for age, race, sex, arrival mode, and ambulatory status on admission, late t-PA was not associated with increased odds of complications or in-hospital deaths (OR, 0.89; 95% CI, 0.49-1.62). The risks and benefits of expansion of the t-PA time window in stroke patients merit further investigation. Key Words: Acute stroke—thrombolytics—health services. Ó 2012 by National Stroke Association
Among acute ischemic stroke patients, the early administration of tissue plasminogen activator therapy (t-PA) is associated with increased likelihood of favorable
From the Department of Epidemiology, Gillings School of Global Public Health, University of North Carolina at Chapel Hill, Chapel Hill. Received July 10, 2010; revision received October 13, 2010; accepted December 1, 2010. Address correspondence to Emily C. O’Brien, MSPH, Department of Epidemiology, Campus Box #8050, 137 E Franklin St, Ste 303B, The University of North Carolina at Chapel Hill, Chapel Hill, NC 27514-3526. E-mail: [email protected]. 1052-3057/$ - see front matter Ó 2012 by National Stroke Association doi:10.1016/j.jstrokecerebrovasdis.2010.12.004
functional status after discharge.1-3 The time window during which t-PA may be safely and effectively administered is an important topic in stroke research. Clinical trial data provide evidence for an interaction between time of t-PA administration and postdischarge neurologic outcomes, with earlier administration being associated with an increased likelihood of favorable measures of neurologic status after discharge.2-5 The 2007 American Heart Association (AHA)/American Stroke Association (ASA) Guidelines for Early Management of Acute Ischemic Stroke recommend that t-PA be administered within 3 hours after symptom onset.6 However, the narrow time window during which t-PA may be administered presents a significant barrier to
Journal of Stroke and Cerebrovascular Diseases, Vol. 21, No. 7 (October), 2012: pp 541-546
541
E.C. O’BRIEN ET AL.
542
its use in clinical practice, because the majority of patients arrive beyond the recommended time window.7-10 Expansion of the time window has been suggested as a strategy for increasing the number of patients eligible for t-PA.1,11 Several recent studies show that the use of t-PA beyond 3 hours may improve functional status and confer little increased risk of complications or mortality. The European Cooperative Acute Stroke Study (ECASS) and the Safe Implementation of Thrombolysis in Stroke Monitoring Study (SITS-MOST) documented similar mortality and risk of complications among patients receiving t-PA within 3 hours of symptom and those receiving t-PA from 3 to 4.5 hours of symptom onset.12,13 Data from a pooled analysis of several large, multicenter trials of t-PA suggest that while t-PA effectiveness does appear to diminish with increasing time since symptom onset, a positive therapeutic effect is still observed up to 6 hours after symptom onset.5 As a result, the AHA/ASA recently issued a scientific advisory recommending administration of t-PA up to 4.5 hours after symptom onset.14 However, confirmation of the safety and efficacy of t-PA beyond the 3-hour time window in observational studies is needed. We examined mean length of hospital stay and risk of hemorrhagic complications and in-hospital mortality among acute stroke patients receiving t-PA within and beyond the 3-hour time window since symptom onset.
Methods Study Population The North Carolina Stroke Care Collaborative (NCSCC) is 1 of 6 Paul Coverdell National Acute Stroke Registries (PCNASR) established to measure, track, and improve the quality of stroke care. The NCSCC comprises 53 hospitals in 39 of 100 counties in North Carolina, representing approximately 61% of all stroke discharges in the state.15 Participating hospitals prospectively enroll presumptive stroke cases and collect data on established quality of care indicators throughout the hospitalization. Trained staff members at each hospital complete a card containing 30 stroke care–related questions for each patient and enter these data into an interactive online database. Additional details of the study design have been published.16,17
Exclusions Between January 2005 and February 2010, 53 participating hospitals enrolled 40,907 patients 18 years of age and older. Of these, 1070 (2.6%) received t-PA during the acute stroke hospitalization. Patients with documented contraindications for t-PA (n 5 5), those who were not admitted for transfer or observation only status (n 5 58), those without complete mortality and covariate information (n 5 28), and those who were missing data that precluded
the calculation of time from last known well to t-PA administration (n 5 93) were excluded from the analysis, for a final study population of 886 patients.
Variable Definitions The main exposure—timing of receipt of t-PA—was defined as ‘‘early’’ (receipt #3 hours since documented time last known well) or ‘‘late’’ (receipt.3 hours since time last known well). The main outcome was any adverse clinical outcome, defined to include in-hospital deaths or hemorrhagic complications. In-hospital deaths were identified using uniform billing form (UB-92) codes or from recorded information indicating that the patient had died during hospitalization. Hemorrhagic complications were defined as a documented symptomatic intracranial hemorrhage or life-threatening, serious systemic hemorrhage within 36 hours of t-PA receipt. Because of the small number of adverse events, death and hemorrhagic complications were combined into a composite endpoint. Length of hospital stay (in days) was calculated using the hospital admission date and the date of discharge or death. We included several clinically relevant demographic and event variables in our analyses, including age (,45, 45-65, 65-80, and $80 years), gender (male or female), race (white or nonwhite), arrival mode (transport by emergency medical services [EMS] or private transport), and ambulatory status at admission (able to ambulate independently or unable to ambulate independently). We also examined in-hospital delay (time from arrival to t-PA administration) and prehospital delay (time from symptom onset to hospital arrival) between early and late t-PA patients.
Statistical Methods and Analysis We examined differences in participant characteristics by early and late receipt of t-PA using c2 tests for independence for categorical variables and t tests for continuous variables. We used multivariable binomial logistic regression to estimate odds ratios (ORs) and 95% confidence intervals (CIs) for the associations between adverse clinical outcomes and timing of t-PA receipt. Multivariable linear regression models were used to estimate mean length of hospital stay for early and late t-PA patients. Because length of stay was not normally distributed, it was log-transformed in the regression analyses. Effect modification of timing of receipt of tPA by age, race, and gender were tested using likelihood ratio tests with an a priori significance level of a 5 0.10. We examined these associations using crude, minimally adjusted (age, race, and sex) and fully adjusted (age, race, sex, arrival status, and ambulatory status on admission) regression models. Because NCSCC registry patients are clustered within hospitals, there was potential for underestimation of
CLINICAL OUTCOMES AND RECEIPT OF t-PA BEYOND 3 HOURS
standard errors. We repeated selected analyses using a multilevel modeling approach that accounted for the nesting of patients within hospitals. Our conclusions remained unchanged; therefore, we present the original models without the multilevel modeling approach. SAS software was used for all analyses (v 9.1; SAS Institute, Cary, NC).
Results Of 886 eligible patients who received t-PA, 781 (88.2%) were administered t-PA within 3 hours of the time that they were last known well (‘‘early’’ t-PA), and 105 (11.9%) were administered t-PA 3 to 6 hours after the time last known well (‘‘late’’ t-PA). Clinical characteristics of stroke patients who received t-PA by time of administration are presented in Table 1. Gender, race, and age distribution were similar among early and late t-PA patients. Early t-PA patients were more likely to arrive by EMS
543
than were late t-PA patients (86.3% v 77.1%; P 5 .01). Adverse events occurred in 16.3% (127/781) of early t-PA patients and 14.3% (15/105) of late t-PA patients. The proportion of patients experiencing hemorrhagic complications within 36 hours of receiving t-PA was similar among the early and late t-PA patients (7.0% v 6.7%; P 5 .89), as was the proportion dying death during the hospitalization (12.0% v 10.5%; P 5 .64). Early t-PA patients had shorter mean prehospital delays than late t-PA patients (59 minutes v 101 minutes; P , .001) and shorter in-hospital delay times (78.4 minutes v 114.3 minutes; P , .001). Unadjusted mean length of hospital stay was longer for early t-PA patients than for late t-PA patients (5.3 days v 4.2 days; P 5 .004). The results of the logistic regression analysis of late t-PA and adverse clinical outcomes are shown in Table 2. Late t-PA administration was associated with a lower odds of an adverse event in crude models, but the estimate is imprecise as shown by the wide CI (OR,
Table 1. Clinical characteristics of stroke patients by receipt of tissue plasminogen activator: The North Carolina Stroke Care Collaborative (2005-2010)
Variables Mean age (95% CI) Gender Male Female Race White Nonwhite Arrival mode EMS Non-EMS Medical history Hypertension Diabetes Smoking Stroke/TIA Ambulation status on admission Ambulating Not ambulating Complicationsz No Yes Died on discharge No Yes Mean in-hospital delay time, min 95% CI Mean prehospital delay time, min 95% CI Mean length of hospital stay, days 95% CI
Received t-PA* (n 5 886; 2.2%)
Did not receive t-PA (n 5 39,685; 97.8%)
P valuey
68.0 (67.1-69.1)
68.3 (68.1-68.4)
.72
437 (49.3) 449 (50.7)
18531 (46.7) 21154 (53.3)
.12
653 (73.7) 233 (26.3)
28327 (71.4) 11358 (28.6)
.13
755 (85.2) 131 (14.8)
21730 (54.8) 17955 (45.2)
,.0001
634 (71.6) 223 (25.2) 194 (21.9) 186 (21.0)
29162 (73.5) 12390 (31.2) 8556 (21.6) 11374 (28.7)
.20 ,.0001 .81 ,.0001
51 (5.8) 835 (94.2)
4926 (12.4) 34759 (87.6)
,.0001
824 (93.0) 62 (7.0) 781 (88.1) 105 (11.9) 78.4 (75.1-81.6) 58.9 (56.9-61.0) 5.3 (5.0-5.6)
— — 36793 (92.7) 2838 (7.3) — 277.7 (272.9-282.5) 3.55 (3.52-3.58)
,.0001 ,.0001 ,.0001 ,.0001
Abbreviations: CI, confidence interval; EMS, emergency medical services; TIA, transient ischemic attack; t-PA, tissue plasminogen activator. *Intravenous tissue plasminogen activator initiated at this hospital. yTwo-sided c2 test of equal proportions or Student t test of different means. zComplications included symptomatic intracranial hemorrhage or serious systemic hemorrhage.
E.C. O’BRIEN ET AL.
544
Table 2. Clinical characteristics of stroke patients who received tissue plasminogen activator by time of tissue plasminogen activator administration: The North Carolina Stroke Care Collaborative (2005-2010) Variables
Early* (n 5 781, 88.2%)
Late (n 5 105, 11.9%)
P valuey
Mean age (95% CI) Gender Male Female Race White Nonwhite Arrival mode EMS Non-EMS Medical history Hypertension Diabetes Smoking Stroke/TIA Ambulation status on admission Ambulating Not ambulating Complicationsz No Yes Died on discharge No Yes Mean in-hospital delay time, min 95% CI Mean prehospital delay time, min 95% CI Mean length of hospital stay, days 95% CI
68.1 (67.1-69.1)
67.3 (64.3-70.3)
.61
392 (50.2) 389 (49.8)
45 (42.9) 60 (57.1)
.16
574 (73.5) 207 (26.5)
79 (75.2) 26 (24.8)
.70
674 (86.3) 107 (13.7)
81 (77.1) 24 (22.9)
.01
562 (72.0) 193 (24.7) 177 (22.7) 164 (21.0)
72 (68.6) 30 (28.6) 17 (16.2) 22 (21.0)
.47 .39 .13 .99
46 (5.9) 735 (94.1)
5 (4.8) 100 (95.2)
.64
726 (93.0) 55 (7.0)
98 (93.3) 7 (6.7)
.89
687 (88.0) 94 (12.0) 78.4 (75.1-81.6) 58.9 (56.9-61.0) 5.3 (5.0-5.6)
94 (89.5) 11 (10.5) 114.3 (105.2-123.7) 101.3 (92.0-110.7) 4.2 (3.6-4.8)
.64 ,.001 ,.001 .004
Abbreviations: CI, confidence interval; EMS, emergency medical services; TIA, transient ischemic attack; t-PA, tissue plasminogen activator. *Early tissue plasminogen activator patients were administered tissue plasminogen activator within 3 hours of the documented last known well time. yTwo-sided c2 test of equal proportions or Student t test of different means. zComplications included symptomatic intracranial hemorrhage or serious systemic hemorrhage.
0.86; 95% CI, 0.48-1.53). Similar patterns were observed in models controlling for age, race, and sex and in models controlling for age, race, sex, arrival mode, and ambulation status on admission (Table 3). When analyses were repeated, restricting the upper limit of the window for late t-PA patients to 4.5 hours after symptom onset, this association did not persist in adjusted models (OR, 1.00; 95% CI, 0.54-1.82; Table 4). The results of the linear regression analysis modeling mean length of hospital stay are shown in Figure 1. Unadjusted mean length of stay (days) was shorter for late t-PA patients (OR, 4.2; 95% CI, 3.6-4.9) than for early t-PA patients (OR, 5.3; 95% CI, 5.0-5.6). After controlling for age, race, and sex, length of stay remained shorter for late (OR, 3.9; 95% CI, 3.0-4.9) than for early t-PA patients (OR, 4.8; 95% CI, 3.9-5.8), although this difference was no longer statistically significant. Results were similar after controlling for age, race, sex, arrival mode, and ambulation status on admission.
Discussion Among NCSCC acute stroke patients, receipt of t-PA more than 3 hours after the time when last known well was not associated with an increased risk of in-hospital mortality or hemorrhagic complications. In addition, we did not observe significant differences in mean adjusted length of stay between early and late t-PA patients. We included all patients who received t-PA from 3 to 6 hours after symptom onset. While one study reported a benefit of t-PA up to 6 hours after symptom onset, the recent recommendations from the AHA/ASA limit the time window to 4.5 hours after symptom onset.5,14 A smaller but not significant proportion of patients with late t-PA administration had adverse events than did those receiving t-PA within a 3-hour time window. However, in a secondary logistic regression analysis excluding 9 patients who received t-PA from 4.5 to 6 hours after time last known well, this association disappeared.
CLINICAL OUTCOMES AND RECEIPT OF t-PA BEYOND 3 HOURS
545
Table 3. Association between tissue plasminogen activator administered beyond 3 hours* and adverse eventsy among all patients receiving tissue plasminogen activator more than 3 hours since symptom onset by model type in the North Carolina Stroke Care Collaborative (2005-2010) Odds ratio (95% CI)
Model description Model 1: Crude Model 2: Model 1 1 age, race, and sex Model 3: Model 2 1 arrival modez and ambulation status on admission
Early t-PA (reference)
Late t-PA
1.00 1.00
0.86 (0.48-1.53) 0.86 (0.48-1.55)
1.00
0.89 (0.49-1.62)
Abbreviations: CI, confidence interval; t-PA, tissue plasminogen activator. *Tissue plasminogen activator administered beyond 3 hours of the documented last known well time. yAdverse events included in-hospital death or documented complications (symptomatic intracranial hemorrhage or serious systemic hemorrhage). zArrival mode defined as emergency medical services or non– emergency medical services.
The 2009 AHA/ASA recommendations for extension of the t-PA time window identify a specific group of exclusion criteria for t-PA receipt beyond 3 hours after time last known well, including a history of both diabetes and stroke, age $ 80 years, and current oral anticoagulant use before admission.14 To investigate the relationship Table 4. Association between tissue plasminogen activator administered beyond 3 hours* and adverse eventsy among patients receiving tissue plasminogen activator from 3 to 4.5 hours since symptom onset by model type in the North Carolina Stroke Care Collaborative (2005-2010) Odds ratio (95% CI)
Model description Model 1: Crude Model 2: Model 1 1 age, race, and sex Model 3: Model 2 1 arrival modez and ambulation status on admission
Early t-PA (reference)
Late t-PA
1.00 1.00
0.95 (0.53-1.71) 0.96 (0.53-1.74)
1.00
1.00 (0.54-1.82)
*Tissue plasminogen activator administered beyond 3 hours of the documented last known well time. yAdverse events included in-hospital death or documented complications (symptomatic intracranial hemorrhage or serious systemic hemorrhage). zArrival mode defined as emergency medical services or non– emergency medical services.
Figure 1. Length of stay in days by time of tissue plasminogen activator (t-PA) administration and model type in the North Carolina Stroke Care Collaborative (2005-2010).
between these exclusion criteria and the timing of t-PA administration, we stratified patients into 2 groups: those with at least 1 of the exclusion criteria, and those without any of the exclusion criteria. The estimated risk of adverse events by timing of administration did not vary across these strata. Confounding by indication is an important concern in pharmacologic intervention research. Because the study population was limited to patients who ultimately received t-PA, this concern was diminished. It is possible that differences in clinical history or sociodemographic characteristics between the early and late t-PA patients may have introduced selection bias. One study suggests that physicians are more likely to administer t-PA to patients who arrive beyond the recommended time window if they have fewer comorbidities, are younger, or are experiencing a less severe event.13 While we did not have information on stroke severity, we did examine a potential interaction between age and the timing of t-PA administration and found that time of t-PA administration did not vary by patient age. In addition, in supplemental analyses, the inclusion of a history of hypertension, diabetes, and stroke in logistic regression models did not change effect estimates (data not shown). Hospital characteristics may affect t-PA use and in-hospital mortality in acute stroke patients. Because of the availability of specialists and increased resources to treat complications, academic medical centers and certified primary stroke centers (PSC) may be better equipped to administer t-PA to patients arriving outside the recommended time window.18-20 However, the inclusion of terms for teaching status and PSC status in the models did not change our results. The results of the length of stay analysis showed slightly longer mean hospital stays among early t-PA patients than among late t-PA patients. In supplemental analyses that were restricted to the subset of patients
546
who were alive at discharge, we observed similar patterns. However, the CIs for these estimates were wide, and a t test for comparing the means of the early and late groups was not statistically significant. We also examined length of stay among patients who ultimately died; however, the numbers of deaths in the early versus late treatment groups were small (11 and 94, respectively). We did not follow patients after discharge, and an analysis of differences in median days until in-hospital death was not statistically significant. This study has several limitations. First, 86 patients were missing the time when last known well or did not have a documented time of t-PA administration. These patients had similar distributions of all included covariates that were similar to patients with nonmissing times; however, we cannot rule that the 2 groups differ on other nonmeasured covariates. In addition, because the NCCSC does not follow patients after hospital discharge, we were not able to examine longer-term outcomes, such as functional status at 3 months. Therefore we could not examine differences in the efficacy of t-PA at improving functionality in acute stroke patients after discharge. Longitudinal cohort studies with the ability to examine the association between late t-PA and long-term functional status are needed. Finally, our relatively small sample size limits our ability to detect statistically significant differences between risk of death or complications and mean length of hospital stay between the time groups. We did not detect differences in the proportion patients experiencing hemorrhagic complications and in-hospital deaths by timing of the receipt of t-PA. Additional population-based studies on treatment patterns, while not able to provide level I evidence for treatment impact like clinical trials, would be useful to better understand the patterns of use and outcomes currently achieved in community-based settings.
References 1. Mori E, Yoneda Y, Tabuchi M, et al. Intravenous recombinant tissue plasminogen activator in acute carotid artery territory stroke. Neurology 1992;42:976-982. 2. Tissue plasminogen activator for acute ischemic stroke. The National Institute of Neurological Disorders and Stroke rt-PA Stroke Study Group. N Engl J Med 1995; 333:1581-1587. 3. Hacke W, Kaste M, Fieschi C, et al. Intravenous thrombolysis with recombinant tissue plasminogen activator for acute hemispheric stroke. The European Cooperative Acute Stroke Study (ECASS). JAMA 1995;274:1017-1025. 4. Hacke W, Kaste M, Fieschi C, et al. Randomised doubleblind placebo-controlled trial of thrombolytic therapy with intravenous alteplase in acute ischaemic stroke (ECASS II). Second European-Australasian Acute Stroke Study Investigators. Lancet 1998;352:1245-1251. 5. Hacke W, Donnan G, Fieschi C, et al. Association of outcome with early stroke treatment: pooled analysis of ATLANTIS, ECASS, and NINDS rt-PA stroke trials. Lancet 2004;363:768-774.
E.C. O’BRIEN ET AL. 6. Adams HP Jr, del Zoppo G, Alberts MJ, et al. Guidelines for the early management of adults with ischemic stroke: a guideline from the American Heart Association/ American Stroke Association Stroke Council, Clinical Cardiology Council, Cardiovascular Radiology and Intervention Council, and the Atherosclerotic Peripheral Vascular Disease and Quality of Care Outcomes in Research Interdisciplinary Working Groups: The American Academy of Neurology affirms the value of this guideline as an educational tool for neurologists. Circulation 2007;115:e478-e534. 7. Kwan J, Hand P, Sandercock P. A systematic review of barriers to delivery of thrombolysis for acute stroke. Age Ageing 2004;33:116-121. 8. Barr J, McKinley S, O’Brien E, et al. Patient recognition of and response to symptoms of TIA or stroke. Neuroepidemiology 2006;26:168-175. 9. Moser DK, Kimble LP, Alberts MJ, et al. Reducing delay in seeking treatment by patients with acute coronary syndrome and stroke: A scientific statement from the American Heart Association Council on cardiovascular nursing and stroke council. Circulation 2006;114:168-182. 10. Rose KM, Rosamond WD, Huston SL, et al. Predictors of time from hospital arrival to initial brain-imaging among suspected stroke patients: The North Carolina Collaborative Stroke Registry. Stroke 2008;39:3262-3267. 11. del Zoppo GJ, Poeck K, Pessin MS, et al. Recombinant tissue plasminogen activator in acute thrombotic and embolic stroke. Ann Neurol 1992;32:78-86. 12. Hacke W, Kaste M, Bluhmki E, et al. Thrombolysis with alteplase 3 to 4.5 hours after acute ischemic stroke. N Engl J Med 2008;359:1317-1329. 13. Wahlgren N, Ahmed N, Davalos A, et al. Thrombolysis with alteplase 3-4.5 h after acute ischaemic stroke (SITS-ISTR): An observational study. Lancet 2008;372: 1303-1309. 14. Del Zoppo GJ, Saver JL, Jauch EC, et al. Expansion of the time window for treatment of acute ischemic stroke with intravenous tissue plasminogen activator: A science advisory from the American Heart Association/American Stroke Association. Stroke 2009;40:2945-2948. 15. Demographics NCS. Revised County Population Estimates for July 1, 2005, North Carolina State Demographics, Federal State Cooperative Program for Population Estimates (FSCPE). Raleigh, North Carolina. 2005. Available at: www.sdc.state.nc.us/. 16. Wattigney WA, Croft JB, Mensah GA, et al. Establishing data elements for the Paul Coverdell National Acute Stroke Registry: Part 1: Proceedings of an expert panel. Stroke 2003;34:151-156. 17. Centers for Disease Control and Prevention. Prehospital and hospital delays after stroke onset—United States, 2005-2006. MMWR Morb Mortal Wkly Rep 2007;56: 474-478. 18. Lattimore SU, Chalela J, Davis L, et al. Impact of establishing a primary stroke center at a community hospital on the use of thrombolytic therapy: The NINDS Suburban Hospital Stroke Center experience. Stroke 2003;34:e55-e57. 19. Meretoja A, Roine RO, Kaste M, et al. Effectiveness of primary and comprehensive stroke centers: PERFECT stroke: A nationwide observational study from Finland. Stroke 2010;41:1102-1107. 20. Jeng JS, Tang SC, Deng IC, et al. Stroke center characteristics which influence the administration of thrombolytic therapy for acute ischemic stroke: A national survey of stroke centers in Taiwan. J Neurol Sci 2009;281:24-27.
Background: Tissue plasminogen activator therapy (t-PA) is associated with improved neurologic outcomes and reduced disability from ischemic stroke. The current guidelines stipulate that patients receive t-PA within 3 hours of symptom onset. However, actual practice patterns vary, and little is known about patient outcomes when t-PA is received outside of the recommended time window. Methods: We examined mean length of hospital stay, t-PA–related complications, and in-hospital death by time of t-PA administration in North Carolina Stroke Care Collaborative (NCSCC) patients. The NCSCC includes 53 hospitals that enroll patients presenting with stroke-like symptoms. Of 40,907 patients enrolled between January 2005 and February 2010, 1070 (2.6%) received t-PA. Of these, 88.2% received t-PA within 3 hours of symptom onset (‘‘early’’) and 30.3% received t-PA between 3 and 6 hours after symptom onset (‘‘late’’). Results: Unadjusted mean length of stay (days) was longer among early patients (5.0 days; 95% confidence interval [CI], 4.7-5.3) than late patients (3.6 days; 95% CI, 3.1-4.2). t-PA–related complications were similar among early (7.0%; 55/781) and late patients (6.7%; 7/102; P 5.89). The proportion of in-hospital deaths was similar among late (10.5%) and early patients (12.0%). We used multivariable logistic regression to estimate odds ratios (ORs) and 95% CIs for the associations between late t-PA status and patient outcomes. Conclusions: In models controlling for age, race, sex, arrival mode, and ambulatory status on admission, late t-PA was not associated with increased odds of complications or in-hospital deaths (OR, 0.89; 95% CI, 0.49-1.62). The risks and benefits of expansion of the t-PA time window in stroke patients merit further investigation. Key Words: Acute stroke—thrombolytics—health services. Ó 2012 by National Stroke Association
Among acute ischemic stroke patients, the early administration of tissue plasminogen activator therapy (t-PA) is associated with increased likelihood of favorable
From the Department of Epidemiology, Gillings School of Global Public Health, University of North Carolina at Chapel Hill, Chapel Hill. Received July 10, 2010; revision received October 13, 2010; accepted December 1, 2010. Address correspondence to Emily C. O’Brien, MSPH, Department of Epidemiology, Campus Box #8050, 137 E Franklin St, Ste 303B, The University of North Carolina at Chapel Hill, Chapel Hill, NC 27514-3526. E-mail: [email protected]. 1052-3057/$ - see front matter Ó 2012 by National Stroke Association doi:10.1016/j.jstrokecerebrovasdis.2010.12.004
functional status after discharge.1-3 The time window during which t-PA may be safely and effectively administered is an important topic in stroke research. Clinical trial data provide evidence for an interaction between time of t-PA administration and postdischarge neurologic outcomes, with earlier administration being associated with an increased likelihood of favorable measures of neurologic status after discharge.2-5 The 2007 American Heart Association (AHA)/American Stroke Association (ASA) Guidelines for Early Management of Acute Ischemic Stroke recommend that t-PA be administered within 3 hours after symptom onset.6 However, the narrow time window during which t-PA may be administered presents a significant barrier to
Journal of Stroke and Cerebrovascular Diseases, Vol. 21, No. 7 (October), 2012: pp 541-546
541
E.C. O’BRIEN ET AL.
542
its use in clinical practice, because the majority of patients arrive beyond the recommended time window.7-10 Expansion of the time window has been suggested as a strategy for increasing the number of patients eligible for t-PA.1,11 Several recent studies show that the use of t-PA beyond 3 hours may improve functional status and confer little increased risk of complications or mortality. The European Cooperative Acute Stroke Study (ECASS) and the Safe Implementation of Thrombolysis in Stroke Monitoring Study (SITS-MOST) documented similar mortality and risk of complications among patients receiving t-PA within 3 hours of symptom and those receiving t-PA from 3 to 4.5 hours of symptom onset.12,13 Data from a pooled analysis of several large, multicenter trials of t-PA suggest that while t-PA effectiveness does appear to diminish with increasing time since symptom onset, a positive therapeutic effect is still observed up to 6 hours after symptom onset.5 As a result, the AHA/ASA recently issued a scientific advisory recommending administration of t-PA up to 4.5 hours after symptom onset.14 However, confirmation of the safety and efficacy of t-PA beyond the 3-hour time window in observational studies is needed. We examined mean length of hospital stay and risk of hemorrhagic complications and in-hospital mortality among acute stroke patients receiving t-PA within and beyond the 3-hour time window since symptom onset.
Methods Study Population The North Carolina Stroke Care Collaborative (NCSCC) is 1 of 6 Paul Coverdell National Acute Stroke Registries (PCNASR) established to measure, track, and improve the quality of stroke care. The NCSCC comprises 53 hospitals in 39 of 100 counties in North Carolina, representing approximately 61% of all stroke discharges in the state.15 Participating hospitals prospectively enroll presumptive stroke cases and collect data on established quality of care indicators throughout the hospitalization. Trained staff members at each hospital complete a card containing 30 stroke care–related questions for each patient and enter these data into an interactive online database. Additional details of the study design have been published.16,17
Exclusions Between January 2005 and February 2010, 53 participating hospitals enrolled 40,907 patients 18 years of age and older. Of these, 1070 (2.6%) received t-PA during the acute stroke hospitalization. Patients with documented contraindications for t-PA (n 5 5), those who were not admitted for transfer or observation only status (n 5 58), those without complete mortality and covariate information (n 5 28), and those who were missing data that precluded
the calculation of time from last known well to t-PA administration (n 5 93) were excluded from the analysis, for a final study population of 886 patients.
Variable Definitions The main exposure—timing of receipt of t-PA—was defined as ‘‘early’’ (receipt #3 hours since documented time last known well) or ‘‘late’’ (receipt.3 hours since time last known well). The main outcome was any adverse clinical outcome, defined to include in-hospital deaths or hemorrhagic complications. In-hospital deaths were identified using uniform billing form (UB-92) codes or from recorded information indicating that the patient had died during hospitalization. Hemorrhagic complications were defined as a documented symptomatic intracranial hemorrhage or life-threatening, serious systemic hemorrhage within 36 hours of t-PA receipt. Because of the small number of adverse events, death and hemorrhagic complications were combined into a composite endpoint. Length of hospital stay (in days) was calculated using the hospital admission date and the date of discharge or death. We included several clinically relevant demographic and event variables in our analyses, including age (,45, 45-65, 65-80, and $80 years), gender (male or female), race (white or nonwhite), arrival mode (transport by emergency medical services [EMS] or private transport), and ambulatory status at admission (able to ambulate independently or unable to ambulate independently). We also examined in-hospital delay (time from arrival to t-PA administration) and prehospital delay (time from symptom onset to hospital arrival) between early and late t-PA patients.
Statistical Methods and Analysis We examined differences in participant characteristics by early and late receipt of t-PA using c2 tests for independence for categorical variables and t tests for continuous variables. We used multivariable binomial logistic regression to estimate odds ratios (ORs) and 95% confidence intervals (CIs) for the associations between adverse clinical outcomes and timing of t-PA receipt. Multivariable linear regression models were used to estimate mean length of hospital stay for early and late t-PA patients. Because length of stay was not normally distributed, it was log-transformed in the regression analyses. Effect modification of timing of receipt of tPA by age, race, and gender were tested using likelihood ratio tests with an a priori significance level of a 5 0.10. We examined these associations using crude, minimally adjusted (age, race, and sex) and fully adjusted (age, race, sex, arrival status, and ambulatory status on admission) regression models. Because NCSCC registry patients are clustered within hospitals, there was potential for underestimation of
CLINICAL OUTCOMES AND RECEIPT OF t-PA BEYOND 3 HOURS
standard errors. We repeated selected analyses using a multilevel modeling approach that accounted for the nesting of patients within hospitals. Our conclusions remained unchanged; therefore, we present the original models without the multilevel modeling approach. SAS software was used for all analyses (v 9.1; SAS Institute, Cary, NC).
Results Of 886 eligible patients who received t-PA, 781 (88.2%) were administered t-PA within 3 hours of the time that they were last known well (‘‘early’’ t-PA), and 105 (11.9%) were administered t-PA 3 to 6 hours after the time last known well (‘‘late’’ t-PA). Clinical characteristics of stroke patients who received t-PA by time of administration are presented in Table 1. Gender, race, and age distribution were similar among early and late t-PA patients. Early t-PA patients were more likely to arrive by EMS
543
than were late t-PA patients (86.3% v 77.1%; P 5 .01). Adverse events occurred in 16.3% (127/781) of early t-PA patients and 14.3% (15/105) of late t-PA patients. The proportion of patients experiencing hemorrhagic complications within 36 hours of receiving t-PA was similar among the early and late t-PA patients (7.0% v 6.7%; P 5 .89), as was the proportion dying death during the hospitalization (12.0% v 10.5%; P 5 .64). Early t-PA patients had shorter mean prehospital delays than late t-PA patients (59 minutes v 101 minutes; P , .001) and shorter in-hospital delay times (78.4 minutes v 114.3 minutes; P , .001). Unadjusted mean length of hospital stay was longer for early t-PA patients than for late t-PA patients (5.3 days v 4.2 days; P 5 .004). The results of the logistic regression analysis of late t-PA and adverse clinical outcomes are shown in Table 2. Late t-PA administration was associated with a lower odds of an adverse event in crude models, but the estimate is imprecise as shown by the wide CI (OR,
Table 1. Clinical characteristics of stroke patients by receipt of tissue plasminogen activator: The North Carolina Stroke Care Collaborative (2005-2010)
Variables Mean age (95% CI) Gender Male Female Race White Nonwhite Arrival mode EMS Non-EMS Medical history Hypertension Diabetes Smoking Stroke/TIA Ambulation status on admission Ambulating Not ambulating Complicationsz No Yes Died on discharge No Yes Mean in-hospital delay time, min 95% CI Mean prehospital delay time, min 95% CI Mean length of hospital stay, days 95% CI
Received t-PA* (n 5 886; 2.2%)
Did not receive t-PA (n 5 39,685; 97.8%)
P valuey
68.0 (67.1-69.1)
68.3 (68.1-68.4)
.72
437 (49.3) 449 (50.7)
18531 (46.7) 21154 (53.3)
.12
653 (73.7) 233 (26.3)
28327 (71.4) 11358 (28.6)
.13
755 (85.2) 131 (14.8)
21730 (54.8) 17955 (45.2)
,.0001
634 (71.6) 223 (25.2) 194 (21.9) 186 (21.0)
29162 (73.5) 12390 (31.2) 8556 (21.6) 11374 (28.7)
.20 ,.0001 .81 ,.0001
51 (5.8) 835 (94.2)
4926 (12.4) 34759 (87.6)
,.0001
824 (93.0) 62 (7.0) 781 (88.1) 105 (11.9) 78.4 (75.1-81.6) 58.9 (56.9-61.0) 5.3 (5.0-5.6)
— — 36793 (92.7) 2838 (7.3) — 277.7 (272.9-282.5) 3.55 (3.52-3.58)
,.0001 ,.0001 ,.0001 ,.0001
Abbreviations: CI, confidence interval; EMS, emergency medical services; TIA, transient ischemic attack; t-PA, tissue plasminogen activator. *Intravenous tissue plasminogen activator initiated at this hospital. yTwo-sided c2 test of equal proportions or Student t test of different means. zComplications included symptomatic intracranial hemorrhage or serious systemic hemorrhage.
E.C. O’BRIEN ET AL.
544
Table 2. Clinical characteristics of stroke patients who received tissue plasminogen activator by time of tissue plasminogen activator administration: The North Carolina Stroke Care Collaborative (2005-2010) Variables
Early* (n 5 781, 88.2%)
Late (n 5 105, 11.9%)
P valuey
Mean age (95% CI) Gender Male Female Race White Nonwhite Arrival mode EMS Non-EMS Medical history Hypertension Diabetes Smoking Stroke/TIA Ambulation status on admission Ambulating Not ambulating Complicationsz No Yes Died on discharge No Yes Mean in-hospital delay time, min 95% CI Mean prehospital delay time, min 95% CI Mean length of hospital stay, days 95% CI
68.1 (67.1-69.1)
67.3 (64.3-70.3)
.61
392 (50.2) 389 (49.8)
45 (42.9) 60 (57.1)
.16
574 (73.5) 207 (26.5)
79 (75.2) 26 (24.8)
.70
674 (86.3) 107 (13.7)
81 (77.1) 24 (22.9)
.01
562 (72.0) 193 (24.7) 177 (22.7) 164 (21.0)
72 (68.6) 30 (28.6) 17 (16.2) 22 (21.0)
.47 .39 .13 .99
46 (5.9) 735 (94.1)
5 (4.8) 100 (95.2)
.64
726 (93.0) 55 (7.0)
98 (93.3) 7 (6.7)
.89
687 (88.0) 94 (12.0) 78.4 (75.1-81.6) 58.9 (56.9-61.0) 5.3 (5.0-5.6)
94 (89.5) 11 (10.5) 114.3 (105.2-123.7) 101.3 (92.0-110.7) 4.2 (3.6-4.8)
.64 ,.001 ,.001 .004
Abbreviations: CI, confidence interval; EMS, emergency medical services; TIA, transient ischemic attack; t-PA, tissue plasminogen activator. *Early tissue plasminogen activator patients were administered tissue plasminogen activator within 3 hours of the documented last known well time. yTwo-sided c2 test of equal proportions or Student t test of different means. zComplications included symptomatic intracranial hemorrhage or serious systemic hemorrhage.
0.86; 95% CI, 0.48-1.53). Similar patterns were observed in models controlling for age, race, and sex and in models controlling for age, race, sex, arrival mode, and ambulation status on admission (Table 3). When analyses were repeated, restricting the upper limit of the window for late t-PA patients to 4.5 hours after symptom onset, this association did not persist in adjusted models (OR, 1.00; 95% CI, 0.54-1.82; Table 4). The results of the linear regression analysis modeling mean length of hospital stay are shown in Figure 1. Unadjusted mean length of stay (days) was shorter for late t-PA patients (OR, 4.2; 95% CI, 3.6-4.9) than for early t-PA patients (OR, 5.3; 95% CI, 5.0-5.6). After controlling for age, race, and sex, length of stay remained shorter for late (OR, 3.9; 95% CI, 3.0-4.9) than for early t-PA patients (OR, 4.8; 95% CI, 3.9-5.8), although this difference was no longer statistically significant. Results were similar after controlling for age, race, sex, arrival mode, and ambulation status on admission.
Discussion Among NCSCC acute stroke patients, receipt of t-PA more than 3 hours after the time when last known well was not associated with an increased risk of in-hospital mortality or hemorrhagic complications. In addition, we did not observe significant differences in mean adjusted length of stay between early and late t-PA patients. We included all patients who received t-PA from 3 to 6 hours after symptom onset. While one study reported a benefit of t-PA up to 6 hours after symptom onset, the recent recommendations from the AHA/ASA limit the time window to 4.5 hours after symptom onset.5,14 A smaller but not significant proportion of patients with late t-PA administration had adverse events than did those receiving t-PA within a 3-hour time window. However, in a secondary logistic regression analysis excluding 9 patients who received t-PA from 4.5 to 6 hours after time last known well, this association disappeared.
CLINICAL OUTCOMES AND RECEIPT OF t-PA BEYOND 3 HOURS
545
Table 3. Association between tissue plasminogen activator administered beyond 3 hours* and adverse eventsy among all patients receiving tissue plasminogen activator more than 3 hours since symptom onset by model type in the North Carolina Stroke Care Collaborative (2005-2010) Odds ratio (95% CI)
Model description Model 1: Crude Model 2: Model 1 1 age, race, and sex Model 3: Model 2 1 arrival modez and ambulation status on admission
Early t-PA (reference)
Late t-PA
1.00 1.00
0.86 (0.48-1.53) 0.86 (0.48-1.55)
1.00
0.89 (0.49-1.62)
Abbreviations: CI, confidence interval; t-PA, tissue plasminogen activator. *Tissue plasminogen activator administered beyond 3 hours of the documented last known well time. yAdverse events included in-hospital death or documented complications (symptomatic intracranial hemorrhage or serious systemic hemorrhage). zArrival mode defined as emergency medical services or non– emergency medical services.
The 2009 AHA/ASA recommendations for extension of the t-PA time window identify a specific group of exclusion criteria for t-PA receipt beyond 3 hours after time last known well, including a history of both diabetes and stroke, age $ 80 years, and current oral anticoagulant use before admission.14 To investigate the relationship Table 4. Association between tissue plasminogen activator administered beyond 3 hours* and adverse eventsy among patients receiving tissue plasminogen activator from 3 to 4.5 hours since symptom onset by model type in the North Carolina Stroke Care Collaborative (2005-2010) Odds ratio (95% CI)
Model description Model 1: Crude Model 2: Model 1 1 age, race, and sex Model 3: Model 2 1 arrival modez and ambulation status on admission
Early t-PA (reference)
Late t-PA
1.00 1.00
0.95 (0.53-1.71) 0.96 (0.53-1.74)
1.00
1.00 (0.54-1.82)
*Tissue plasminogen activator administered beyond 3 hours of the documented last known well time. yAdverse events included in-hospital death or documented complications (symptomatic intracranial hemorrhage or serious systemic hemorrhage). zArrival mode defined as emergency medical services or non– emergency medical services.
Figure 1. Length of stay in days by time of tissue plasminogen activator (t-PA) administration and model type in the North Carolina Stroke Care Collaborative (2005-2010).
between these exclusion criteria and the timing of t-PA administration, we stratified patients into 2 groups: those with at least 1 of the exclusion criteria, and those without any of the exclusion criteria. The estimated risk of adverse events by timing of administration did not vary across these strata. Confounding by indication is an important concern in pharmacologic intervention research. Because the study population was limited to patients who ultimately received t-PA, this concern was diminished. It is possible that differences in clinical history or sociodemographic characteristics between the early and late t-PA patients may have introduced selection bias. One study suggests that physicians are more likely to administer t-PA to patients who arrive beyond the recommended time window if they have fewer comorbidities, are younger, or are experiencing a less severe event.13 While we did not have information on stroke severity, we did examine a potential interaction between age and the timing of t-PA administration and found that time of t-PA administration did not vary by patient age. In addition, in supplemental analyses, the inclusion of a history of hypertension, diabetes, and stroke in logistic regression models did not change effect estimates (data not shown). Hospital characteristics may affect t-PA use and in-hospital mortality in acute stroke patients. Because of the availability of specialists and increased resources to treat complications, academic medical centers and certified primary stroke centers (PSC) may be better equipped to administer t-PA to patients arriving outside the recommended time window.18-20 However, the inclusion of terms for teaching status and PSC status in the models did not change our results. The results of the length of stay analysis showed slightly longer mean hospital stays among early t-PA patients than among late t-PA patients. In supplemental analyses that were restricted to the subset of patients
546
who were alive at discharge, we observed similar patterns. However, the CIs for these estimates were wide, and a t test for comparing the means of the early and late groups was not statistically significant. We also examined length of stay among patients who ultimately died; however, the numbers of deaths in the early versus late treatment groups were small (11 and 94, respectively). We did not follow patients after discharge, and an analysis of differences in median days until in-hospital death was not statistically significant. This study has several limitations. First, 86 patients were missing the time when last known well or did not have a documented time of t-PA administration. These patients had similar distributions of all included covariates that were similar to patients with nonmissing times; however, we cannot rule that the 2 groups differ on other nonmeasured covariates. In addition, because the NCCSC does not follow patients after hospital discharge, we were not able to examine longer-term outcomes, such as functional status at 3 months. Therefore we could not examine differences in the efficacy of t-PA at improving functionality in acute stroke patients after discharge. Longitudinal cohort studies with the ability to examine the association between late t-PA and long-term functional status are needed. Finally, our relatively small sample size limits our ability to detect statistically significant differences between risk of death or complications and mean length of hospital stay between the time groups. We did not detect differences in the proportion patients experiencing hemorrhagic complications and in-hospital deaths by timing of the receipt of t-PA. Additional population-based studies on treatment patterns, while not able to provide level I evidence for treatment impact like clinical trials, would be useful to better understand the patterns of use and outcomes currently achieved in community-based settings.
References 1. Mori E, Yoneda Y, Tabuchi M, et al. Intravenous recombinant tissue plasminogen activator in acute carotid artery territory stroke. Neurology 1992;42:976-982. 2. Tissue plasminogen activator for acute ischemic stroke. The National Institute of Neurological Disorders and Stroke rt-PA Stroke Study Group. N Engl J Med 1995; 333:1581-1587. 3. Hacke W, Kaste M, Fieschi C, et al. Intravenous thrombolysis with recombinant tissue plasminogen activator for acute hemispheric stroke. The European Cooperative Acute Stroke Study (ECASS). JAMA 1995;274:1017-1025. 4. Hacke W, Kaste M, Fieschi C, et al. Randomised doubleblind placebo-controlled trial of thrombolytic therapy with intravenous alteplase in acute ischaemic stroke (ECASS II). Second European-Australasian Acute Stroke Study Investigators. Lancet 1998;352:1245-1251. 5. Hacke W, Donnan G, Fieschi C, et al. Association of outcome with early stroke treatment: pooled analysis of ATLANTIS, ECASS, and NINDS rt-PA stroke trials. Lancet 2004;363:768-774.
E.C. O’BRIEN ET AL. 6. Adams HP Jr, del Zoppo G, Alberts MJ, et al. Guidelines for the early management of adults with ischemic stroke: a guideline from the American Heart Association/ American Stroke Association Stroke Council, Clinical Cardiology Council, Cardiovascular Radiology and Intervention Council, and the Atherosclerotic Peripheral Vascular Disease and Quality of Care Outcomes in Research Interdisciplinary Working Groups: The American Academy of Neurology affirms the value of this guideline as an educational tool for neurologists. Circulation 2007;115:e478-e534. 7. Kwan J, Hand P, Sandercock P. A systematic review of barriers to delivery of thrombolysis for acute stroke. Age Ageing 2004;33:116-121. 8. Barr J, McKinley S, O’Brien E, et al. Patient recognition of and response to symptoms of TIA or stroke. Neuroepidemiology 2006;26:168-175. 9. Moser DK, Kimble LP, Alberts MJ, et al. Reducing delay in seeking treatment by patients with acute coronary syndrome and stroke: A scientific statement from the American Heart Association Council on cardiovascular nursing and stroke council. Circulation 2006;114:168-182. 10. Rose KM, Rosamond WD, Huston SL, et al. Predictors of time from hospital arrival to initial brain-imaging among suspected stroke patients: The North Carolina Collaborative Stroke Registry. Stroke 2008;39:3262-3267. 11. del Zoppo GJ, Poeck K, Pessin MS, et al. Recombinant tissue plasminogen activator in acute thrombotic and embolic stroke. Ann Neurol 1992;32:78-86. 12. Hacke W, Kaste M, Bluhmki E, et al. Thrombolysis with alteplase 3 to 4.5 hours after acute ischemic stroke. N Engl J Med 2008;359:1317-1329. 13. Wahlgren N, Ahmed N, Davalos A, et al. Thrombolysis with alteplase 3-4.5 h after acute ischaemic stroke (SITS-ISTR): An observational study. Lancet 2008;372: 1303-1309. 14. Del Zoppo GJ, Saver JL, Jauch EC, et al. Expansion of the time window for treatment of acute ischemic stroke with intravenous tissue plasminogen activator: A science advisory from the American Heart Association/American Stroke Association. Stroke 2009;40:2945-2948. 15. Demographics NCS. Revised County Population Estimates for July 1, 2005, North Carolina State Demographics, Federal State Cooperative Program for Population Estimates (FSCPE). Raleigh, North Carolina. 2005. Available at: www.sdc.state.nc.us/. 16. Wattigney WA, Croft JB, Mensah GA, et al. Establishing data elements for the Paul Coverdell National Acute Stroke Registry: Part 1: Proceedings of an expert panel. Stroke 2003;34:151-156. 17. Centers for Disease Control and Prevention. Prehospital and hospital delays after stroke onset—United States, 2005-2006. MMWR Morb Mortal Wkly Rep 2007;56: 474-478. 18. Lattimore SU, Chalela J, Davis L, et al. Impact of establishing a primary stroke center at a community hospital on the use of thrombolytic therapy: The NINDS Suburban Hospital Stroke Center experience. Stroke 2003;34:e55-e57. 19. Meretoja A, Roine RO, Kaste M, et al. Effectiveness of primary and comprehensive stroke centers: PERFECT stroke: A nationwide observational study from Finland. Stroke 2010;41:1102-1107. 20. Jeng JS, Tang SC, Deng IC, et al. Stroke center characteristics which influence the administration of thrombolytic therapy for acute ischemic stroke: A national survey of stroke centers in Taiwan. J Neurol Sci 2009;281:24-27.