Abstract
Background
The benefit of intravenous thrombolytic therapy in acute brain ischemia is strongly time dependent.
Methods
The Get with the Guidelines-Stroke (GWTG-Stroke) database was analyzed to characterize ischemic stroke patients arriving to hospital Emergency Departments (EDs) within 60 minutes of last known well time from 4/1/2003-12/30/2007.
Results
During the 4.75 year study period, among 253,148 ischemic stroke patients arriving directly by ambulance or private vehicle to 905 hospital EDs, 106,924 (42.2%) had documented exact last known well times. Onset to door time was ≤ 60 minutes in 30,220 (28.3%), 61-180 minutes in 33,858 (31.7%), and >180 minutes in 42,846 (40.1%). Features most strongly distinguishing ≤ 60, 61-180, and > 180 minutes arriving patients were: greater stroke severity (median NIHSS 8.0 vs 6.0 vs 4.0, p <.0001) and more frequent arrival by ambulance (79.0%. vs 72.2% vs 55.0%, p <.0001). Compared with 61-180 minute arrivers, golden hour patients received IV thrombolytic therapy more frequently (27.1% vs 12.9%, OR 2.51, 95% CI 2.41-2.61, p <.0001), but door to needle time (DTN) was longer (mean 90.6 vs 76.7 minutes, p <.0001). DTN ≤ 60 minutes was achieved in 18.3% of golden hour patients.
Conclusions
At GWTG-Stroke hospital EDs, more than one quarter of patients with documented onset time, and at least one eighth of all ischemic stroke patients, arrive within 1 hour of onset, where they receive thrombolytic therapy more frequently but more slowly than late arrivers. These findings support public health initiates to increase early presentation and shorten door to needle times in patients arriving within the “golden hour.”
The benefit of intravenous thrombolytic therapy in acute brain ischemia is strongly time dependent. Therapeutic yield is maximal in the first minutes after symptom onset and declines rapidly over the next 4.5 hours.1, 2 In the typical large artery ischemic stroke, in each minute in which reperfusion is delayed, 2 million nerve cells die.3 Among every 100 patients treated with intravenous therapy, for every 10 minute delay in the start of lytic infusion within the 1-3 hour treatment time period, 1 fewer patient has an improved disability outcome.2 Consequently, patients who present to hospital within the first 60 minutes of onset have the greatest opportunity to benefit from recanalization therapy. Because of the critical importance of rapid treatment, national recommendations for hospitals that accept acute stroke patients in their emergency department are to complete the clinical and imaging evaluation of the patient and initiate lytic therapy within 1 hour of patient arrival.4 The Joint Commission target for Primary Stroke Centers is to achieve a door to needle time of within 60 minutes in 80% or more of patients.
Originally developed in the setting of trauma treatment, the term the “golden hour” is now a general concept in emergency medicine that is applied to conditions in which hyperacute therapy is more effective than later intervention, including trauma, myocardial ischemia, septic shock, cardiopulmonary resuscitation, and stroke. The frequency, characteristics, and treatment of ischemic stroke patients arriving at hospitals within the golden hour nationally have not previously been well-characterized. Several large registry studies in the United States and internationally have provided important information regarding patients arriving in the under three hour window for lytic therapy.5-9 Of concern, cohort studies have suggested an inverse relationship between time from symptom onset to hospital arrival and door to needle time among stroke patients treated with intravenous tissue plasminogen activator (TPA).10 Among patients arriving between 100-130 minutes after onset, rapid ED care occurred and door to needle times under or near 60 minutes were often attained, permitting therapy start within the 180 minute limit of the drug label. In contrast, among patients arriving early after onset, door to needle times were often extended, and therapy frequently still not administered until nearly the three hour mark. The representativeness of these small treatment cohorts is not known.
The Get With the Guidelines–Stroke national dataset offers an opportunity to examine the presenting characteristics of ischemic stroke patients arriving within the golden hour, factors associated with early presentation, the rapidity of lytic care initiation, and explore determinants of efficient lytic care in a large, nationally representative cohort.
Methods
The American Heart Association and American Stroke Association (AHA/ASA) launched the GWTG-Stroke initiative focused on the redesign of hospital systems of care to improve the quality of care of patients with stroke and transient ischemic attack.11, 12 GWTG uses a Web-based patient management tool (PMT, Outcome, Cambridge, MA) to collect clinical data on consecutively admitted patients, provide decision support, and enable real-time online reporting features. Following an initial pilot phase conducted in 8 states, the GWTG-Stroke program was made available in April 2003 to any hospital in the US.13 Data from hospitals that joined the program anytime between April 2003 and Dec 2007 were included in this analysis. Each participating hospital received either human research approval to enroll cases without individual patient consent under the common rule, or a waiver of authorization and exemption from subsequent review by their Institutional Review Board. Outcome Sciences, Inc. serves as the data collection and coordination center for GWTG. The Duke Clinical Research Institute serves as the data analysis center and has an agreement to analyze the aggregate de-identified data for research purposes.
Case Identification and Data Abstraction
Trained hospital personnel were instructed to ascertain consecutive acute stroke admissions by either prospective clinical identification, retrospective identification using International Classification of Diseases (ICD)-9 discharge codes, or a combination. Methods used for prospective identification varied, but included regular surveillance of emergency department records (i.e., presenting symptoms and chief complaints), ward census logs, and/or neurological consultations. The eligibility of each acute stroke admission was confirmed at chart review prior to abstraction. Patient data abstracted using the PMT included demographics, medical history, initial head computerized tomography findings, in-hospital treatment and events, discharge treatment, mortality, and discharge destination. Data on hospital-level characteristics (i.e., bed size, academic or non-academic status, annual volume of stroke discharges, and geographical region) were collected from the American Hospital Association.14
For this study, the GWTG-Stroke database was analyzed to characterize ischemic stroke patients arriving to hospital emergency departments (EDs) ≤ 60 minutes compared to > 60 minutes after symptom onset, during the time period 4/1/2003-12/30/2007. Patient level variables were analyzed for all patients meeting study entry criteria. Hospital level determinants were analyzed for hospitals that entered five or more patients meeting study criteria in the database. Contingency tables were generated to explore group differences in demographics (age, sex), stroke severity, arrival mode (ambulance, private vehicle), door to needle (DTN) time, door to imaging (DTI) time, and outcome destination at discharge. Chi-square test for nominal data and Kruskal-Wallis tests for ordinal and continuous data were used as tests for unadjusted statistical associations. Statistical significance was defined as P≤0.01. Generalized estimating equations logistic regression models, accounting for within-hospital clustering, were generated to identify independent predictors of onset to door (OTD) time ≤ 60 minutes and of DTN ≤ 60 minutes. General details of the candidate patient and hospital variables and the modeling process have been previously described.11 Statistical significance was defined as P≤0.01. All statistical analyses were performed using SAS Version 9.1 software (SAS Institute, Cary, NC).
Results
During the 4.75 year time period, at 905 hospital sites, data for 431,170 ischemic stroke and TIA patients were entered into the GWTG-Stroke database. The main analyses of this study were performed upon the 106,924 patients in this cohort with ischemic stroke, a documented last known well time, and presentation directly to the ED by ambulance or private vehicle. Among excluded patients were: 74,671 patients who did not present directly to the ED (including in hospital stroke, elective admission directly to floor, secondary transfer from another hospital); 103,351 ED arriving patients with final diagnoses of transient ischemic attack; and 146,224 direct ED-arriving ischemic stroke patients in whom last known well time (LKWT) was not documented.
Patient and hospital level characteristics among patients with and without a documented last known well time are shown in Table 1. Large differences were noted in arrival by EMS and use of TPA (both higher in documented LKWT patients) and modest differences in other features, including stroke severity (higher in documented LKWT patients) and race (lower frequency of blacks in documented LKWT patients).
Table 1.
Patient- and Hospital-Level Characteristics of Ischemic Stroke Patients With and Without Documented Last Known Well Time (LKWT)
| LKWT Documented | LKWT Not Documented | P value | |
|---|---|---|---|
| N | 106,924 | 146,224 | |
| Patient Level Characteristics | |||
| Age | 74 (14.35) | 75 (14.39) | <.0001 |
| Sex (Female) | 51.5% | 54.6% | <.0001 |
| Race-ethnicity | |||
| White, non-Hispanic | 75.4% | 72.1% | <.0001 |
| Black | 13.4% | 16.7% | |
| Asian | 2.3% | 2.3% | |
| Arrival by EMS (vs private transport) | 67.2% | 54.7% | <.0001 |
| NIHSS* (median, IQR) | 6 (2-13) | 4 (1-9) | <.0001 |
| Hx of atrial fib/flutter | 20.2% | 16.8% | <.0001 |
| Prior stroke/TIA | 30.5% | 31.5% | <.0001 |
| CAD/prior MI | 28.0% | 27.4% | <.0005 |
| Carotid Stenosis | 4.3% | 4.4% | <.4445 |
| Peripheral vascular dz | 4.8% | 5.3% | <.0001 |
| Diabetes Mellitus | 27.5% | 31.5% | <.0001 |
| Hx of HTN | 73.6% | 74.7% | <.0001 |
| Smoker | 17.1% | 17.2% | <.4879 |
| Hx of dyslipidemia | 36.2% | 34.2% | <.0001 |
| IV TPA treatment | 11.7% | 0.5% | <.0001 |
| Hospital Level Characteristics | |||
| IS-TIA Admissions per year | |||
| 301+ | 31.2% | 30.1% | <.0001 |
| 101-300 | 57.5% | 57.8% | |
| 0-100 | 11.3% | 12.2% | |
| Hospital Size (beds) | 375 | 367 | <.0001 |
| Hospital Type (Non-Academic) | 38.5% | 39.9% | <.0001 |
| Hospital Region | |||
| West | 19.3% | 16.7% | <.0001 |
| South | 35.3% | 38.1% | |
| Midwest | 19.7% | 19.4% | |
| Northeast | 25.7% | 25.9% | |
Reasons for Not Know include: symptom onset time not valid or not documented, hospital arrival time not valid or not documented, symptom onset documented as after hospital arrival time, or no documentation present.
NIHSS values were recorded in 148,681 patients, 58.71% of the cohort.
Among the direct ED-arriving ischemic stroke patients with last known well time documented, OTD time was 60 minutes or less in 30,220 (28.3%), 61-180 minutes in 33,858 (31.7%), and >180 minutes in 42,846 (40.1%). Among the subgroup who arrived within 60 minutes, mean OTD time was 39.9 minutes (SD 14.8). In the most recent study year, 2007, among the 809 facilities contributing data, GWTG-Stroke hospitals cared for 10,497 golden-hour arriving ischemic stroke patients.
Table 2 shows patient- and hospital level characteristics of three ischemic stroke time-of-arrival cohorts. All groups were similar in age and gender. In race-ethnicity, both 1 hour or under and 1-3 hour arriving patients, compared with beyond 3 hour arriving patients, were slightly more often non-Hispanic white and less often black or Asian. Among those patients in whom stroke severity was documented (n=51,738), severity was greatest among golden hour arriving patients (median NIHSS 8), intermediate among 1-3 hour patients (NIHSS 6) and least among beyond 3 hour patients (NIHSS 4). A similar graded difference was observed in the frequency of arrival at hospital by ambulance, which occurred in 79.0% of 1 hour or under patients, 72.2% of 1-3 hour patients, and 55.0% of greater than 3 hour patients. Considering hospital characteristics, arrival within the golden hour occurred mildly more often at hospitals located in the Northeast and the West.
Table 2.
Patient- and Hospital-Level Characteristics of Ischemic Stroke Patients Arriving in Different Time Windows
| ≤60 minutes | 61-180 minutes | >180 minutes | P value | |
|---|---|---|---|---|
| N | 30220 | 33858 | 42846 | |
| Patient Level Characteristics | ||||
| Age | 71.3 (14.4) | 72.0 (14.3) | 70.6 (14.2) | < 0.0001 |
| Sex (Female) | 50.8% | 52.2% | 51.5% | 0.002 |
| Race-ethnicity | ||||
| White, non-Hispanic | 77.3% | 77.5% | 72.5% | < 0.0001 |
| Black | 11.8% | 11.9% | 15.8% | |
| Asian | 2.0% | 2.1% | 2.7% | |
| Arrival by EMS (vs private transport) | 79.0% | 72.2%% | 55.0.% | < 0.0001 |
| NIHSS* (median, IQR) | 8 (3-16) | 6 (2-12) | 4 (2-9) | < 0.0001 |
| H/o atrial fib/flutter | 24.3% | 21.7% | 16.2% | < 0.0001 |
| Prior stroke/TIA | 30.0% | 32.0% | 29.6% | < 0.0001 |
| CAD/prior MI | 29.4% | 28.9% | 26.3% | < 0.0001 |
| Carotid Stenosis | 4.2% | 4.4% | 4.4% | 0.57 |
| Peripheral vascular dz | 4.7% | 5.0% | 4.8% | 0.32 |
| Diabetes Mellitus | 23.4% | 27.0% | 30.8% | < 0.0001 |
| H/o HTN | 71.9% | 73.7% | 74.9% | < 0.0001 |
| Smoker | 84.4% | 84.6% | 80.6% | < 0.0001 |
| H/o dyslipidemia | 35.1% | 36.5% | 36.8% | < 0.0001 |
| Hospital Level Characteristics | ||||
| Annual Volume of IS-TIA Admissions | ||||
| 301+ | 29.5% | 32.3% | 31.6% | < 0.0001 |
| 101-300 | 58.3% | 56.7% | 57.6% | |
| 0-100 | 12.2% | 11.0% | 10.8% | |
| Hospital Size (beds) | 358 | 380 | 380 | < 0.0001 |
| Hospital Type (Non-Academic) | 41.0% | 38.0% | 37.2% | < 0.0001 |
| Hospital Region | ||||
| West | 20.4% | 18.4% | 19.2% | < 0.0001 |
| South | 34.2% | 36.3% | 35.5% | |
| Midwest | 19.3% | 19.1% | 20.4% | |
| Northeast | 26.1% | 26.2% | 24.9% | |
NIHSS values were recorded in 51,378 patients, 48.1% of the cohort.
Patient and hospital factors independently associated with symptom onset to door times less than or equal to 1 hour are shown in Table 3. The most powerful characteristics independently associated with increased odds of early arrival were severe neurologic deficit, arrival by ambulance rather than private transport, and atrial fibrillation. Patient factors associated with decreased odds of arrival in the golden hour included older age, female sex, and atherosclerotic risk factors (HTN, tobacco, diabetes). Hospital factors associated with decreasing odds of golden hour arrival included hospital location in the South (US census region definition) and higher annual number of stroke admissions.
Table 3.
Patient and Hospital Level Characteristics Independently Associated with ED Arrival within the First 60 Minutes of Onset
| Characteristic | OR (95% CI) | P value |
|---|---|---|
| Severe deficit (NIHSS 9-41 vs 0-3) | 1.84 (1.76-1.93) | <0.001 |
| Arrival mode (EMS vs Private transportation | 1.78 (1.70-1.87) | <0.001 |
| H/o atrial fibrillation | 1.21 (1.16-1.26) | <0.001 |
| Moderate deficit (NIHSS 4-8 vs 0-3) | 1.16 (1.10-1.22) | <0.001 |
| CAD/prior MI | 1.08 (1.03-1.12) | <0.001 |
| Prior stroke/TIA | 0.96 (0.92-1.00) | 0.049 |
| H/o HTN | 0.95 (0.91-0.99) | 0.018 |
| Sex (F vs M) | 0.94 (0.90-0.98) | 0.002 |
| Age (per 10 year increase) | 0.91 (0.90-0.92) | <0.001 |
| Race-ethnicity, Black vs White, non-Hispanic | 0.91 (0.86-0.97 | 0.004 |
| Hospital Region (South vs West) | 0.87 (0.78-0.98) | 0.024 |
| Moderate annual hospital stroke admits (101-300 vs ≤100) | 0.87 (0.78-0.97) | 0.012 |
| Smoker | 0.84 (0.80-0.88) | <0.001 |
| Race-ethnicity, Asian vs White, non-Hispanic | 0.78 (0.68-0.89) | <0.001 |
| Diabetes Mellitus | 0.77 (0.74-0.80) | <0.001 |
| High annual hospital stroke admits (>300 vs ≤100) | 0.76 (0.66-0.87) | <0.001 |
Table reflects modeling performed with 55,057 patients with full data available, including NIHSS. No major differences (apart from NIHSS) were observed when model was constructed using an additional 50.962 patients without recorded NIHSS.
Over the course of the study period, intravenous TPA was administered in 12,545 of the study cohort ischemic stroke patients. (In addition, TPA was administered to 159 direct ED-arriving patients who received a final diagnosis of transient ischemic attack.) The 12,545 TPA-treated ischemic stroke patients represented 11.8% of all direct ED-arriving ischemic stroke patients with documented last known well times and 5.0% of all direct ED-arriving ischemic stroke patients. Among the ischemic stroke patients receiving IV TPA, 8111 (64.7%) arrived at hospital within the first 60 minutes, 4327 (34.5%) between 61-180 minutes, and 107 (0.9%) beyond 180 minutes. Compared with 61-180 minute arrivers, patients arriving in the first 60 minutes received IV thrombolytic therapy more frequently (27.1% vs 12.9%, unadjusted OR 2.51, 95% CI 2.41-2.61, p <.0001).
Door to needle time across all IV TPA treated patients was mean 86 minutes (SD 41.6). An inverse relationship between OTD and DTN time was noted, with correlation coefficient -0.30 (Figure 1). DTN time was longer among patients arriving in the first hour then in patients arriving in hours 1-3 (mean 90.6 vs. 76.7 minutes, p <.0001). The distribution of DTN times in 1 hour or under arriving patients is shown in Figure 2. Total elapsed time from symptom onset to treatment in patients arriving within the first hour was mean 129 minutes (SD 39). Among these early arrival patients, 1.6% received TPA within 60 minutes of onset, 11.0% between 61-90 minutes, 30.2% between 91-120 minutes, 31.5% between 121-150 minutes, 21.7% between 151-180 minutes, and 4.0% later than 180 minutes.
Figure 1.
Relation of onset to door versus door to needle times among all IV TPA within 3 hours treated ischemic stroke patients (n=11883). Mean OTD time was 56.3 (± 28.5) and DTN time 84.1 (± 29.0). The correlation coefficient was -0.30.
Figure 2.
Distribution of door to needle times among patients arriving in ED within first 60 minutes after onset. Bracket shows proportion treated within DTN ≤ 60 minutes stretch target.
The target DTN time of ≤ 60 minutes was achieved in 18.3% of golden hour-arriving patients. Patient characteristics in the DTN ≤ 60 minute patients and > 60 minute groups among the golden hour arriving cohort are shown in Table 4. DTN ≤ 60 minute patients were slightly younger and more often male compared to > 60 minute patients. In contrast, stroke deficit severity, ambulance mode of ED arrival, and race did not differ among the DTN ≤ 60 minute and > 60 minute patients. The proportion of patients with a DTN time of ≤60 minutes increased modestly over time, from 12.8% in 2003 to 19.5% in 2007, with trendline showing an increase of 1.2% per year. In contrast, there was no relation of achievement of DTN ≤ 60 minutes to the duration of hospital participation in the GWTG-Stroke Program. The proportion of patients with DTN ≤ 60 minutes nominally increased from 18.2% in year 1 to 18.9% in year 5 of GWTG-Stroke participation, with a correlation coefficient of 0.11 (p=0.65).
Table 4.
Frequency of patient and hospital characteristics among golden-hour arriving patients treated with IV TPA under and over 60 minutes after arrival
| DTN ≤ 60 min (n=1425) | DTN > 60 min (N=6345) | P value | |
|---|---|---|---|
| Age | 67.6 (±14.7) | 69.8 (±14.7) | <0.0001 |
| Sex (% female) | 43.7% | 49.9% | <0.0001 |
| Race-ethnicity | |||
| White, non-Hispanic | 1102 (77.3%) | 4909 (77.4%) | 0.2607 |
| Black | 152 (10.7%) | 766 (12.1%) | |
| Asian | 35 (2.5%) | 142 (2.2%) | |
| Ambulance arrival | 1264 (88.7%) | 5549 (87.5%) | 0.1955 |
| NIHSS (median, IQR)* | 14.0 (9.0-18.0) | 13.0 (8.0-19.0) | 0.0461 |
| Calendar year | |||
| 2003 | 29 (12.8%) | 197 (87.2%) | 0.0274 |
| 2004 | 89 (15.9%) | 472 (84.1%) | |
| 2005 | 277 (18.9%) | 1191 (81.1%) | |
| 2006 | 479 (17.8%) | 2210 (82.2%) | |
| 2007 | 551 (19.5%) | 2275 (80.5%) | |
| GWTG-Stroke Year | |||
| 1 | 427 (18.6%) | 1872 (81.4%) | 0.7804 |
| 2 | 420 (17.5%) | 1975 (82.5%) | |
| 3 | 335 (19.1%) | 1417 (80.9%) | |
| 4 | 164 (18.8%) | 707 (81.2%) | |
| 5 | 77 (17.8%) | 355 (82.2%) | |
NIHSS values were recorded in 6043 patients, 77.8% of the cohort.
Among the 905 hospitals enrolling any patient in the database during the study time period, 473 entered 5 or more patients who arrived directly to the ED with onset to door time less than or equal to 60 minutes. Among these hospitals, the proportion of golden hour patients with DTN ≤ 60 minutes was 0-20% at 307 hospitals (64.9%), 21-40% at 132 (27.9%), 41-60% at 30 (6.3%), 61-80% at 4 (0.8%), and 81-100% at none. Dividing hospitals into quartiles, at the 121 hospitals with the highest proportion of golden hour patients with door to needle times ≤ 60 minutes, the proportion of golden hour arriving patients treated within the first hour after arrival ranged from 27-80%; in the 116 hospitals of the second quartile hospital group, the proportion ranged from 15-27%; in the 118 hospitals of the third quartile, the proportion ranged from 3-14%; and in the 118 hospitals of the bottom quartile, the proportion ranged from 0-2%. Table 5 shows the characteristics of the hospitals among these different performance groups. Higher number of patients treated with IV TPA annually was the only hospital characteristic associated with a higher proportion of golden hour patients treated within 60 minutes of arrival. Variables that were not predictive included hospital size, total number of stroke patients encountered, academic or nonacademic status, months in the GWTG-Stroke program, and geographic region.
Table 5.
Characteristics of Hospitals* with Different Rates of Achievement of DTN ≤ 60 minutes among Golden Hour ED-arriving patients
| Highest Quartile | Third Quartile | Second Quartile | Lowest Quartile | P value | |
|---|---|---|---|---|---|
| Annual volume of ischemic stroke patients** | 157 (±101) | 175 (±130) | 206 (±123) | 153(±93) | 0.002 |
| Annual volume of all IV TPA ED stroke patients | 8.2 (±7.4) | 7.8 (±6.2) | 7.8 (±4.5) | 4.8 (±6.4) | <.0001 |
| Proportion of Golden Hour-arriving ED TPA patients with DTN ≤ 60 mins | 38.6% (±9.8) | 20.2% (±3.2) | 9.8% (±3.1) | 0.03% (±0.3) | <.0001 |
| Months in GWTG-Stroke | 43.8 (±17.9) | 47.1 (±18.4) | 47.0 (±18.3) | 43.8(±18.8) | 0.250 |
| Hospital Size | 373.9 (±214.3) | 430.7 (±329.0) | 458.7 (±296.0) | 354.8 (±185.1) | 0.034 |
| Non Academic | 43.0% | 41.4% | 33.1% | 49.1% | 0.094 |
| JC Certified PSC | 69.4% | 54.3% | 67.0% | 53.4% | 0.072 |
| Region | |||||
| West | 20.7% | 13.8% | 23.7% | 14.4% | 0.140 |
| South | 28.1% | 37.9% | 28.0% | 36.4% | |
| Midwest | 20.7% | 12.9% | 22.9% | 16.1% | |
| Northeast | 30.6% | 35.3% | 25.4% | 33.1% | |
Among the 473 of 905 hospitals with 5 or more golden hour-arriving patients in the GWTG-Stroke database. Hospitals divided into quartiles based on proportion of Golden Hour-arriving patients with OTD ≤ 60 minutes. Quartile size and ranges: Highest quartile: 121 hospitals, proportion with DTN ≤ 60 minutes 27-80%; Third quartile: 116 hospitals, 15-27%; Second quartile: 118 hospitals, 3-14%; Lowest quartile: 118 hospitals 0-2%.
As measured in calendar year 2007
JC – Joint Commission, PSC – Primary Stroke Center
Results of multivariate modeling identifying patient and hospital level factors independently associated with DTN times ≤ 60 minutes among golden-hour arriving patients are shown in Table 6. Greater stroke severity increased the odds of lytic treatment start within 1 hour of arrival, while older age, female sex, and history of diabetes or prior stroke/TIA decreased the odds.
Table 6.
Patient and Hospital Level Characteristics Independently Associated with Door to Needle Time ≤ 60 Minutes Among Patients Arriving at ED within the First 60 Minutes of Onset
| Characteristic | OR (95% CI) | P value |
|---|---|---|
| Severe deficit (NIHSS 9-41 vs 0-3) | 2.26 (1.45-3.53) | <0.001 |
| Moderate deficit (NIHSS 4-8 vs 0-3) | 1.71 (1.07-2.74) | 0.026 |
| Calendar year (per 1 year increase) | 1.12 (1.04-1.22) | 0.003 |
| Age (per 10 year increase) | 0.92 (0.88-0.95) | <0.001 |
| Sex (F vs M) | 0.85 (0.75-0.96) | 0.010 |
| Diabetes Mellitus | 0.79 (0.66-0.94) | 0.007 |
| Prior stroke/TIA | 0.77 (0.65-0.91) | 0.002 |
Table reflects modeling performed with 6043 patients with full data available, including NIHSS. No major differences (apart from NIHSS) were observed when model was constructed using an additional 1,727 patients without recorded NIHSS.
Discussion
There have been several important, national and multicenter registry studies of early arriving stroke patients,5-9 but this study is the largest and the first to characterize in detail ischemic stroke patients who arrive at hospital within the first 60 minutes after onset, the “golden hour” when the opportunity to save threatened brain tissue by reperfusion is greatest. The sheer size of the golden hour-arriving population is a notable finding of the investigation. Patients arriving in the first hour of onset accounted for more than one in four ischemic stroke patients who presented to Get With the Guidelines–Stroke hospitals with documented last known well times and at least one in eight of all ED-arriving ischemic stroke patients. Using recent estimates for the annual incidence of ischemic stroke in the US and the proportion of ischemic patients admitted to hospital, from these findings it may be projected that more than 55,000 Americans each year present to acute care hospitals within the first 60 minutes of onset of ischemic stroke.
Because early time of presentation is critical to early start of therapy, a public health priority is to increase even further the proportion of acute ischemic stroke patients presenting within the first 60 minutes after onset.7 In the GWTG-Stroke dataset, the two most powerful determinants of arrival in the first 60 minutes were greater severity of stroke deficits on the NIH Stroke Scale and arrival by ambulance rather than private vehicle. These findings suggest that public health messages have a substantial opportunity to increase the proportion of early arriving patients, by educating patients, family members, and on scene bystanders to recognize the symptoms of stroke and to react to less severe as well as more severe deficits by calling 911 and activating the Emergency Medical System (EMS). Another factor affecting arrival in the first hour was race-ethnicity, with blacks and Asians less likely to arrive in the golden hour than non-Hispanic whites. In a recent study in 13 states and the District of Columbia, awareness of stroke warning symptoms and the importance of activating the 911 system was less common among African-American, Hispanic, and Other (predominantly Asian) race-ethnic groups than among whites.15 Several studies have found that white, non-Hispanic patients with stroke are more likely to arrive at the Emergency Department in early time windows and to receive thrombolytic therapy than are African-Americans and other race/ethnic groups.16, 17 These findings suggest a need not only for general public education campaigns, but also for campaigns targeted to distinctive communities, including African-Americans, Hispanics, and Asians. Educational campaigns have increased effectiveness when they are tailored to an individual's cultural heritage.18 Targeted campaigns for stroke awareness would be beneficial, building on the foundation of past and current initiatives, including the American Stroke Association Power to End Stroke campaign, the Beauty Shop Stroke Education Project, “Hip-Hop Stroke”, and the Kids Identifying and Defeating Stroke (KIDS).
Patients who arrived at the Emergency Department within the first 60 minutes had a 2.5-fold higher rate of treatment with intravenous fibrinolytic therapy than patients arriving at 61-180 minutes, with more than 1 in 4 golden hour-arriving patients receiving IV TPA. Prior studies have indicated that, were all ischemic stroke patients to arrive immediately at hospital after onset, about one-quarter would be appropriate candidates for IV recanalization therapy, while three-quarters would have other contraindications to therapy such as having mild strokes, abnormal coagulation studies, or recent surgery.19, 20 Accordingly, it appears that GWTG-Stroke hospitals successfully delivered intravenous fibrinolytic therapy to the great preponderance of patients arriving in the golden hour who were fully eligible for therapy.
However, while the proportion of golden-hour patients treated with fibrinolytic therapy was substantial, the speed of initiation of treatment after hospital arrival was often below the recommended national target of a door-to-needle time of less than or equal to 60 minutes. As in prior, smaller studies,10 an inverse relationship was noted between time of hospital arrival and door to needle time in IV TPA treated stroke patients. Among patients arriving at 100-130 minutes after onset, rapid ED care often occurred and door to needle times under 60 minutes were frequently achieved, permitting therapy start within 180 minutes. However, among patients arriving earlier, door to needle times were often extended. The shorter door-to-needle times in later arriving patients, in part, likely reflects a selection effect. Patients in whom the responding team was unable to start therapy before the 3 hour mark elapsed did not receive treatment and therefore do not enter into analyses of the interval from arrival to therapy start. But this shorter treatment time also likely reflects a systematically more rapid diagnostic and therapeutic response by hospital stroke teams in later arriving patients, when the time remaining to start therapy within the three hour limit was limited. The average DTN time to start in golden-hour arriving patients was more than 1.5 hours and fewer than one in five patients had a door to needle time meeting the national target of under 60 minutes.
It is important to emphasize that the national under 60 minute door-to-needle target was somewhat arbitrarily chosen, promulgated based on the opinion of a small group of experts of what was safely achievable rather than formal time-motion studies or large scale practice experience.4 Multiple studies have found that the target is achieved in only a minority of patients even in very experienced centers. It is perhaps currently best viewed as an ambitious target that centers should over time steadily approach, rather than a minimum target that all centers should currently meet.
Nonetheless, this study identifies substantial opportunities nationally for improvement in the speed of fibrinolytic therapy initiation in acute ischemic stroke patients. Once patients with ischemic stroke have done their part by arriving at a medical center early, it is incumbent upon the receiving hospital to perform rapid diagnostic evaluation and, in appropriate patients, swiftly initiate intravenous fibrinolytic therapy. In golden-hour arriving patients, there are natural human tendencies to use the additional time available before the outermost permitted treatment time (such as 3 hours or 4.5 hours) to increase diagnostic certainty and treatment consensus. Health professionals are naturally tempted to use this time to elicit the history in greater detail, perform a fuller neurologic physical examination, carry out a detailed review of imaging and laboratory tests, discuss more fully the benefits and risks of therapy with patients and on scene family, and reach primary care physicians and other off-scene advisors to arrive at a comfortable, consensus treatment decision. But while there are many good reasons to delay the start of therapy in early arriving patients, they are all trumped by the one, overwhelming reason to hurry – brain is dying all the while that these activities are taking place.
The need to emphasize the door-to-needle time target, rather than a treat before final time window expiration approach, has been further increased by the recent demonstration that IV TPA can confer some modest benefit up to 4.5 hours of onset, with resulting expansion of the treatment window according to European, Canadian, and US national guidelines.21-23 Without the looming hard cutoff of the old 3 hour limit, patients arriving at hospital now in the 2nd hour after onset may be at risk for the slower response and initiation of IV TPA observed in this study for golden-hour patients.
System interventions focused on continuous, iterative quality improvement can reduce door to needle times for ischemic stroke patients. In the two NINDS-TPA trials themselves, the median door to needle time was 64 minutes, even though extensive research informed consents had to be obtained in all patients.24 In regular clinical practice, select centers worldwide have reported mean door to needle times well below 60 minutes, including 25 minutes in Erlangen, Germany,25 29 minutes in Busan, Korea,26 and 38 minutes in Bergen, Norway.27 Successful centers report effective components of programs to accelerate door to needle times include prearrival notification by emergency medical service providers, written protocols for acute triage and patient flow, single call systems to activate all stroke team members, CT or MRI scanner clearance as soon as center is made aware of an incoming patient, storage and rapid access to lytic drug in the Emergency Department, collaboration in developing treatment pathways among physician, nurses, pharmacists, and technologists from Emergency Medicine, Neurology, and Radiology, and continuous data collection to drive iterative system improvement.25-28
Encouraging in our study were observations that achievement of door-to-needle times under 60 minutes was highest at hospitals with a larger volume of IV TPA experience, and a mild temporal improving trend from 2003 to 2007. The number of hospitals with large volume experience is likely to increase in coming years due to several factors, including the increase to 4.5 hours in the time window for IV TPA, the regionalization of emergency stroke care with direct routing of patients to state-designated stroke centers,29 and the emergence into practice of a generation of treatment-oriented neurologists and emergency physicians. The finding that length of time in the GWTG-Stroke program was not associated with an increase in the proportion of patients treated within 60 minutes of arrival suggests a need to revisit and reframe aspects of the GWTG-Stroke toolkit and intervention strategy to highlight the importance of this target and provide concrete strategies for its achievement in various practice settings.
This study has several limitations. Hospitals participating in GWTG-Stroke are likely to have more well-organized stroke systems of care than nonparticipating hospitals, so other US hospitals are likely on average to have worse lytic treatment rates and door to needle times than observed in this cohort. Nonetheless, by the final year of observation, about 23% of US hospitals containing about 41% of licensed US hospital beds were participating in GWTG-Stroke, so this study does reflect a substantial proportion of US practice. The last known well time was documented in 42% of patients. While this rate is higher than in many epidemiologic studies (in which last known well time is often documented only 15-30% of the time), it is lower than desirable. However, is likely that last known well time is more often documented among early-arriving patients, in whom it greatly influences ED management, and less often documented among late-arriving patients, in whom the exact onset time is of less practical importance. In accord with this hypothesis, patients with documented last known well times more often arrived by EMS ambulance and had greater stroke severity, two features associated with earlier arrival. Consequently, the analyses in this study confined to the golden-hour arriving patients likely capture the great preponderance of actual golden-hour arriving patients in the study period. During the study period, a small group of treated patients received a final diagnosis of transient ischemic attack, accounting for 1.3% of the IV TPA treated patients. It may be suggested that the diagnosis in these patients should be reclassified as therapeutically averted strokes and included in the ischemic stroke group. We retained the original GWTG-Stroke database diagnostic categories, so our time to treatment analysis was performed in the 98.7% of IV TPA-treated patients who received a final diagnosis of ischemic stroke. Residual measured and unmeasured confounding variables may have influenced some of the findings.
We investigated the influence of multiple patient and hospital level factors upon care of early arriving patients. However, many additional factors important in fostering rapid care were not captured in the GWTG-Stroke database and therefore not analyzed, including policies of local Emergency Medical Service (EMS) agencies, hospital provision of education programs to EMS, existence of a regional stroke system of care with routing of stroke patients directly to designated stroke centers, location of CT or MRI scanners in the ED, and policies regarding need for ancillary testing before treatment such as coagulation studies, CT angiography and CT perfusion imaging, or multimodal MRI imaging. Data quality is always a concern in registry studies and the GWTG-Stroke registry is implemented by a diverse group of users. To optimize data quality, the GWTG-Stroke program includes detailed training of site chart abstractors, standardized case definitions and coding instructions, predefined logic and range checks on data fields at data entry, audit trails, and regular data quality reports for all sites. Limited source documentation audits at the individual state and site level have shown high data quality and a nationally representative audit is under way. Nevertheless, as in any cardiovascular and stroke registry, data is subject to limitations in the quality and accuracy of the medical records themselves as well as to the quality of medical record abstraction. Further there are a portion of ischemic stroke patients where onset time is unavailable not due to limitations in data quality but because the time of onset cannot be obtained from the patient.
We conclude that golden hour-arriving patients are a substantial population, accounting for one in eight ischemic stroke patients who arrive directly to the Emergency Department. Arrival by ambulance rather than private vehicle was among the most powerful determinants of arrival in the golden hour. Upon arrival, they receive thrombolytic therapy more frequently and earlier than late arrivers. While target door to needle times ≤ 60 minutes are achieved in fewer than one-fifth of golden hour-arriving patients, treatment times show a mild improving national trend over time and are better at high treatment volume centers. These findings support sustained public education efforts to increase the proportion of patients arriving within the first 30-60 minutes after stroke onset by emphasizing the recognition of stroke symptoms and the immediate activation of 911. These data also encourage reinvigorated hospital performance improvement activities to shorten door to needle times in patients who present in the “golden hour” when the volume of salvageable brain and the patient's capacity to benefit from reperfusion therapy is greatest.
Acknowledgments
Funding: Get With the Guidelines-Stroke is funded by the American Heart Association and the American Stroke Association. The program is also supported in part by unrestricted educational grants to the American Heart Association by Pfizer, Inc., New York, NY, and the Merck-Schering Plough Partnership (North Wales, PA), who did not participate in the design, analysis, manuscript preparation or approval. JLS was supported for this work by an American Heart Association PRT Outcomes Research Center Award and by NIH-NINDS Awards P50 NS044378 and U01 NS 44364.
Footnotes
Conflicts of Interest/Disclosures: Dr Saver serves as a member of the Get With the Guidelines Science Subcommittee, as a scientific consultant regarding trial design and conduct to CoAxia, Concentric Medical, Talecris, and Ev3 (all modest); received lecture honoraria from Ferrer and Boehringer Ingelheim (modest); was an unpaid investigator in a multicenter prevention trial sponsored by Boehringer Ingelheim; has declined consulting/honoraria monies from Genentech since 2002; is an employee of the University of California, which holds a patent on retriever devices for stroke.
Dr. Smith receives research support from the NIH (NINDS R01 NS062028) the Canadian Stroke Network, and the Hotchkiss Brain Institute and Canadian Institutes for Health Research, and salary support from the Canadian Institutes for Health Research.
Dr. Fonarow receives research support from the National Institutes of Health (significant), serves as a consultant to Pfizer, Merck, Schering Plough, Bristol Myers Squibb, and Sanofi-Aventis (all modest); receives speaker honoraria from Pfizer, Merck, Schering Plough, Bristol Myers Squibb, and Sanofi-Aventis (all significant); and is an employee of the University of California, which holds a patent on retriever devices for stroke.
Dr. Reeves receives salary support from the Michigan Stroke Registry
Dr. Zhao is a member of the Duke Clinical Research Institute which serves as the AHA GWTG data coordinating center.
Dr. Olson is a member of the Duke Clinical Research Institute which serves as the AHA GWTG data coordinating center.
Dr. Schwamm serves as a consultant to the Research Triangle Institute, CryoCath, and to the Massachusetts Department of Public Health.
References
- 1.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. doi: 10.1056/NEJMoa0804656. [DOI] [PubMed] [Google Scholar]
- 2.Lansberg MG, Schrooten M, Bluhmki E, Thijs VN, Saver JL. Treatment time-specific number needed to treat estimates for tissue plasminogen activator therapy in acute stroke based on shifts over the entire range of the modified Rankin Scale. Stroke. 2009;40:2079–2084. doi: 10.1161/STROKEAHA.108.540708. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Saver JL. Time is brain--quantified. Stroke. 2006;37:263–266. doi: 10.1161/01.STR.0000196957.55928.ab. [DOI] [PubMed] [Google Scholar]
- 4.Marler JR, Winters Jones P, Emr M. The National Institute of Neurological Disorders and Stroke: Proceedings of National Symposium on Rapid Identification and Treatment of Acute Stroke; Bethesda, MD. 1997. [Google Scholar]
- 5.Wahlgren N, Ahmed N, Davalos A, et al. Thrombolysis with alteplase for acute ischaemic stroke in the Safe Implementation of Thrombolysis in Stroke-Monitoring Study (SITS-MOST): an observational study. Lancet. 2007;369:275–282. doi: 10.1016/S0140-6736(07)60149-4. [DOI] [PubMed] [Google Scholar]
- 6.Hill MD, Buchan AM. Thrombolysis for acute ischemic stroke: results of the Canadian Alteplase for Stroke Effectiveness Study. CMAJ. 2005;172:1307–1312. doi: 10.1503/cmaj.1041561. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Evenson KR, Foraker RE, Morris DL, Rosamond WD. A comprehensive review of prehospital and in-hospital delay times in acute stroke care. Int J Stroke. 2009;4:187–199. doi: 10.1111/j.1747-4949.2009.00276.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.George MG, Tong X, McGruder H, et al. Paul Coverdell National Acute Stroke Registry Surveillance - four states, 2005-2007. MMWR Surveill Summ. 2009;58:1–23. [PubMed] [Google Scholar]
- 9.Albers GW, Bates VE, Clark WM, Bell R, Verro P, Hamilton SA. Intravenous tissue-type plasminogen activator for treatment of acute stroke: the Standard Treatment with Alteplase to Reverse Stroke (STARS) study. JAMA. 2000;283:1145–1150. doi: 10.1001/jama.283.9.1145. [DOI] [PubMed] [Google Scholar]
- 10.Albers GW, Bates VE, Clark WM, Bell R, Verro P, Hamilton SA. Intravenous tissue type plasminogen activator for treatment of acute stroke: the Standard Treatment with Alteplase to Reverse Stroke (STARS) study. Journal of the American Medical Association. 2000;283:1145–1150. doi: 10.1001/jama.283.9.1145. [DOI] [PubMed] [Google Scholar]
- 11.Schwamm LH, Fonarow GC, Reeves MJ, et al. Get With the Guidelines-Stroke is associated with sustained improvement in care for patients hospitalized with acute stroke or transient ischemic attack. Circulation. 2009;119:107–115. doi: 10.1161/CIRCULATIONAHA.108.783688. [DOI] [PubMed] [Google Scholar]
- 12.Reeves MJ, Fonarow GC, Zhao X, Smith EE, Schwamm LH. Quality of care in women with ischemic stroke in the GWTG program. Stroke. 2009;40:1127–1133. doi: 10.1161/STROKEAHA.108.543157. [DOI] [PubMed] [Google Scholar]
- 13.LaBresh KA, Reeves MJ, Frankel MR, Albright D, Schwamm LH. Hospital treatment of patients with ischemic stroke or transient ischemic attack using the “Get With The Guidelines” program. Arch Intern Med. 2008;168:411–417. doi: 10.1001/archinternmed.2007.101. [DOI] [PubMed] [Google Scholar]
- 14.American Hospital Association. American Hospital Association Hospital Statistics 2007. Chicago, IL: American Hospital Association; 2007. [Google Scholar]
- 15.Awareness of stroke warning symptoms--13 States and the District of Columbia, 2005. MMWR Morb Mortal Wkly Rep. 2008;57:481–485. [PubMed] [Google Scholar]
- 16.Stansbury JP, Jia H, Williams LS, Vogel WB, Duncan PW. Ethnic Disparities in Stroke: Epidemiology, Acute Care, and Postacute Outcomes. Stroke. 2005;36:374–386. doi: 10.1161/01.STR.0000153065.39325.fd. [DOI] [PubMed] [Google Scholar]
- 17.Schumacher HC, Bateman BT, Boden-Albala B, et al. Use of thrombolysis in acute ischemic stroke: analysis of the Nationwide Inpatient Sample 1999 to 2004. Ann Emerg Med. 2007;50:99–107. doi: 10.1016/j.annemergmed.2007.01.021. [DOI] [PubMed] [Google Scholar]
- 18.Kreuter MW, Wray RJ. Tailored and targeted health communication: strategies for enhancing information relevance. Am J Health Behav. 2003;27 3:S227–232. doi: 10.5993/ajhb.27.1.s3.6. [DOI] [PubMed] [Google Scholar]
- 19.California Acute Stroke Pilot Registry (CASPR) Investigators. Prioritizing interventions to improve rates of thrombolysis for ischemic stroke. Neurology. 2005;64:654–659. doi: 10.1212/01.WNL.0000151850.39648.51. [DOI] [PubMed] [Google Scholar]
- 20.Laloux P, Thijs V, Peeters A, Desfontaines P. Obstacles to the use of intravenous tissue plasminogen activator for acute ischemic stroke. Is time the only barrier? Acta Neurol Belg. 2007;107:103–107. [PubMed] [Google Scholar]
- 21.ESO Executive Committee, ESO Writing Committee. European Stroke Organization Guideline Update - January 2009. [online]. Available at: http://www.eso-stroke.org/pdf/ESO%20Guidelines_update_Jan_2009.pdf.
- 22.Lindsay PBP, Bayley MMD, Hellings CB, Hill MMMD, Woodbury EBMHA, Phillips SM. Canadian best practice recommendations for stroke care (updated 2008) CMAJ. 2008;179:S1–25. [Google Scholar]
- 23.Del Zoppo GJ, Saver JL, Jauch EC, Adams HP., Jr 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. doi: 10.1161/STROKEAHA.109.192535. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.NINDS t-PA Stroke Study Investigators. NINDS t-PA Stroke Study Data Set (on CD-ROM) 2006. [Google Scholar]
- 25.Köhrmannm M, Schellinger P, Breuer L, et al. Avoiding in-hospital delays in early time windows: a standardized stroke thrombolysis workflow in a neurological emergency room. Submitted. [Google Scholar]
- 26.Kim SK, Lee SY, Bae HJ, et al. Pre-hospital notification reduced the door-to-needle time for iv t-PA in acute ischaemic stroke. Eur J Neurol. 2009;16:1331–1335. doi: 10.1111/j.1468-1331.2009.02762.x. [DOI] [PubMed] [Google Scholar]
- 27.Tveiten A, Mygland A, Ljostad U, Thomassen L. Intravenous thrombolysis for ischaemic stroke: short delays and high community-based treatment rates after organisational changes in a previously inexperienced centre. Emerg Med J. 2009;26:324–326. doi: 10.1136/emj.2008.063610. [DOI] [PubMed] [Google Scholar]
- 28.A systems approach to immediate evaluation and management of hyperacute stroke. Experience at eight centers and implications for community practice and patient care The National Institute of Neurological Disorders and Stroke (NINDS) rt-PA Stroke Study Group. Stroke. 1997;28:1530–1540. doi: 10.1161/01.str.28.8.1530. [DOI] [PubMed] [Google Scholar]
- 29.Schwamm LH, Pancioli A, Acker JE, 3rd, et al. Recommendations for the establishment of stroke systems of care: recommendations from the American Stroke Association's Task Force on the Development of Stroke Systems. Stroke. 2005;36:690–703. doi: 10.1161/01.STR.0000158165.42884.4F. [DOI] [PubMed] [Google Scholar]