Results of Intravenous Thrombolysis Within 4.5 to 6 Hours and Updated Results Within 3 to 4.5 Hours of Onset of Acute Ischemic Stroke Recorded in the Safe Implementation of Treatment in Stroke International Stroke Thrombolysis Register (SITS-ISTR): An Observational Study

Importance  Pooled analysis of randomized controlled trials of intravenous thrombolysis shows no statistically significant benefit beyond 4.5 hours, with the possible advantage perhaps offset by risk.

Objective  To compare the outcomes of patients who were treated within 4.5 to 6 hours or within 3 to 4.5 hours of the onset of an ischemic stroke with the outcomes of patients who were treated within 3 hours in the SITS-ISTR.

Design  An observational study based on SITS-ISTR data during the period from 2002 to 2011.

Setting  Acute and emergency care.

Participants  Of 29 618 patients with acute ischemic stroke, 283 (1.0%) were treated within 4.5 to 6 hours of onset, 4056 (13.7%) were within 3 to 4.5 hours of onset, and 25 279 (85.4%) were treated within 3 hours of onset, in compliance with other European Union approval criteria.

Exposure  Intravenous thrombolysis with alteplase.

Main Outcomes and Measures  Functional independence (modified Rankin Scale score of 0-2) and mortality at 3 months and symptomatic intracerebral hemorrhage (SICH). P values are based on comparisons between patients treated within 4.5 to 6 hours or within 3 to 4.5 hours of onset against patients treated within 3 hours of onset.

Results  Results are presented as within 4.5 to 6 hour vs within 3 to 4.5 hours vs within 3 hours. Median time from stroke onset to treatment was 295 vs 210 minutes vs 138 minutes (P < .01), median age was 65 vs 67 vs 68 years (P < .01), and median baseline National Institutes of Health Stroke Scale score was 9 vs 9 vs 12 (P < .01). Rate of functional independence was 61.3% (P = .40) vs 62.7% (P < .01) vs 58.4%; mortality rate was 11.8% (P = .99) vs 11.1% (P = .21) vs 11.8%; and rate of SICH was 2.6% (P = .17) vs 1.8% (P = .27) vs 1.5%. Multivariate analysis detected no significant difference in SICH (P > .05), mortality (P > .05), or independence (P > .05). Time from stroke onset to treatment as a continuous variable was significantly associated with higher rates of SICH and poor 3-month outcome after adjustment for age and National Institutes of Health Stroke Scale score.

Conclusions and Relevance  The treatment remains safe and effective for patients treated within 3 to 4.5 hours compared with patients treated within 3 hours. Our selected group of patients treated within 4.5 to 6 hours of stroke onset did not have worse outcomes than patients treated within 3 hours. An inevitable limitation of our observational study is the possible nonequivalence of the cohorts, particularly the 4.5- to 6-hour cohort relative to the other 2 cohorts.

Intravenous alteplase, a recombinant tissue plasminogen activator, is the only pharmacological therapy that is approved for treatment within 3 hours of ischemic stroke onset. Recently, the treatment approval with alteplase has been extended in many countries to up to 4.5 hours after stroke onset. Most guidelines^1,2 already recommended treatment up to 4.5 hours after stroke onset after the publication of the European Cooperative Acute Stroke Study (ECASS) III trial³ and the Safe Implementation of Treatment in Stroke International Stroke Thrombolysis Register (SITS-ISTR) 4.5-hour study in October 2008.⁴ The pooled analysis⁵ of randomized controlled trials in 2004 and the updated pooled analysis⁶ in 2010 confirmed the benefit of treatment with intravenous alteplase up to 4.5 hours after stroke. There was no statistically significant benefit observed after 4.5 hours; however, 95% CIs for odds ratios (ORs) indicate that a benefit may exist up to 6 hours after stroke onset. An increased mortality rate was observed when treatment was initiated after 4.5 hours.⁶

The SITS-ISTR⁷ is an interactive database over the secure Internet that allows for the registering of unselected patients treated with thrombolysis for acute ischemic stroke. The SITS-ISTR centers are committed to registering all consecutive stroke patients treated with thrombolysis, whether treated on- or off-label.

We aimed to identify patients treated off-label within 4.5 to 6 hours (271-360 minutes) and recorded in the SITS-ISTR and to compare their outcomes with the outcomes of patients treated within 3 hours (≤180 minutes). We also aimed to update the comparison of outcomes between patients treated within 3 to 4.5 hours (181-270 minutes) and patients treated within 3 hours.

In the primary analysis, we included patients with ischemic stroke who were treated with intravenous alteplase (Boehringer Ingelheim) within 6 hours of symptom onset and who were registered in the SITS-ISTR between December 2002 and November 2011 in compliance with other European Union (EU) approval criteria (eTable in Supplement). The main reason for the exclusion of noncompliant treatments was that the inclusion of patients with other protocol violation(s) will make it difficult to interpret the effect of treatment beyond 4.5 hours of stroke symptom onset. Another reason was to compare the present study results with previous results from the main SITS publications,^4,8,9 in which only patients compliant with EU approval criteria were included. In a secondary sensitivity analysis, we included all patients during the same study period. Details of the method, including data collection and management for SITS-ISTR and SITS Monitoring Study (SITS-MOST), have been described previously.⁸

The need for ethical approval or patient consent for participation in the SITS-ISTR varied among participating countries, but approvals were obtained in countries that required this; other countries approved the register for conduct as an anonymized audit. The SITS-MOST was approved by the ethics committee of the Karolinska Institute in Stockholm, Sweden, and by the Swedish Medical Products Agency.

The SITS International Coordination Office performed regular online monitoring of the SITS-ISTR data and checked individual patient data on a regular basis to handle errors or inconsistencies. For a sample of patients included in the SITS-MOST (n = 6483),⁸ source data were verified on-site by monitors under the supervision of their national coordinator.

The main outcomes and measures were functional independence (ie, a modified Rankin Scale [mRS] score of 0-2) and mortality at 3 months and symptomatic intracerebral hemorrhage (SICH) per the SITS-MOST,⁸ which is a local or remote type 2 parenchymal hemorrhage detected on an imaging scan 22 to 36 hours before treatment or earlier if clinically indicated, combined with neurological deterioration of 4 points or more (as measured by the National Institutes of Health Stroke Scale [NIHSS] between baseline and 24 hours) or death.

For comparison with other published work, we also report whether a patient had no or minimal disability (ie, an mRS score of 0-1) at 3 months; SICH per the National Institute of Neurological Disorders and Stroke,¹⁰ which is any ICH detected on any posttreatment imaging scan, combined with any decline in neurologic status as measured by the NIHSS between baseline and 7 days; and SICH per the ECASS II,¹¹ which is any ICH detected on any posttreatment imaging scan, combined with neurological deterioration of 4 points or more (as measured by the NIHSS between baseline and 7 days).

All evaluations of imaging studies and neurological status were performed according to the clinical routine of each hospital. All definitions of SICHs were centrally adjudicated by the SITS International Coordination Office based on the clinical and imaging data entered by the investigators in the registry.

We used descriptive statistics for the baseline and demographic data obtained from the 3 cohorts (ie, patients treated within 4.5-6 hours, patients treated within 3-4.5 hours, and patients treated within 3 hours of onset of ischemic stroke) with regard to the time from stroke onset to treatment. P values are based on comparisons between patients treated within 4.5 to 6 hours or within 3 to 4.5 hours of onset against patients treated within 3 hours of onset. For categorical variables, we calculated percentages by dividing the number of events by the total number of patients, excluding missing or unknown cases. The Pearson χ² test was used for categorical variables, and the t test and the Mann-Whitney U test were used for continuous and ordinal variables, respectively. Odds ratios for the different outcome parameters were calculated by comparing patients treated within 4.5 to 6 hours or within 3 to 4.5 hours of onset with patients treated within 3 hours of onset. We performed multivariate logistic regression analysis after adjusting for age, sex, prestroke mRS score, atrial fibrillation, history of hyperlipidemia, previous stroke earlier than 3 months, antihypertensive therapy, signs of a recent infarct at baseline imaging, and baseline stroke severity as measured by NIHSS score because these variables were statistically significant in the univariate analysis at the 10% level. However, the time from stroke onset to hospital admission was not included in the multivariate analysis owing to the collinearity with the time from hospital admission to treatment, which was retained in the analysis. The time from stroke onset to treatment as a continuous variable was excluded from the multivariate analysis owing to the risk of overestimation because the study cohorts were subgrouped according to this variable. However, in multivariate analyses, we separately examined the association of the time from stroke onset to treatment as a continuous variable with each outcome parameter. We performed all analyses using Statistica software version 10.0 (StatSoft Inc).

From December 2002 to November 2011, a total of 44 311 patients with ischemic stroke were treated with intravenous thrombolysis within 6 hours of symptom onset and were recorded in the SITS-ISTR. A total of 29 618 patients were included in the primary analysis because they were in compliance with other EU approval criteria (eTable in Supplement), except for time from stroke onset to treatment. Of these patients, 283 (1.0%) were treated within 4.5 to 6 hours of stroke onset, 4056 (13.7%) were treated within 3 to 4.5 hours of stroke onset, and 25 279 (85.4%) were treated within 3 hours of stroke onset. Patients treated within 4.5 to 6 hours were registered from 24 countries and 122 clinical centers (range, 1-19 patients), patients treated within 3 to 4.5 hours were registered from 37 countries and 469 clinical centers (range, 1-127 patients), and patients treated within 3 hours were registered from 40 countries and 723 clinical centers (range, 1-802 patients).

The baseline and demographic characteristics of the patients who otherwise were compliant with EU approval criteria are shown in Table 1. Compared with the cohort of patients who were treated within 3 hours of symptom onset, treatment was initiated a median of 157 minutes later after symptom onset in the 4.5- to 6-hour cohort and a median of 72 minutes later after symptom onset in the 3- to 4.5-hour cohort; the median age was 3 years younger in the 4.5- to 6-hour cohort and 1 year younger in the 3- to 4.5-hour cohort; and stroke severity was 3 points lower (as measured by NIHSS score) in the both 4.5- to 6-hour cohort and the 3- to 4.5-hour cohort. The median time delay from hospital admission to initiation of treatment was 38 minutes longer in the 4.5- to 6-hour cohort and 15 minutes longer in the 3- to 4.5-hour cohort compared with the cohort of patients treated within 3 hours. Baseline imaging scans showed a higher frequency of visible recent infarct in the 4.5- to 6-hour cohort (35%; P < .01) and the 3- to 4.5-hour cohort (22%; P < .01) than in the cohort of patients treated within 3 hours (19%). Computed tomography and/or magnetic resonance angiography were more frequently performed for the 4.5- to 6-hour cohort (38%; P < .01) and the 3- to 4.5-hour cohort (25%; P < .01) than for the cohort of patients treated within 3 hours (19%).

Figure 1 shows the proportion of patients by 30-minute time intervals according to time from stroke onset to treatment. In the 4.5- to 6-hour cohort, about 62% of patients were treated within 271 to 300 minutes, and about 86% of patients were treated within 271 to 330 minutes.

Of the 44 311 patients in the secondary analysis, which included other EU approval violations, 542 (1.2%) were treated within 4.5 to 6 hours, 6449 (14.6%) were treated within 3 to 4.5 hours, and 37 320 (84.2%) were treated within 3 hours after stroke onset. The baseline and demographic characteristics of these patients showed a similar pattern to the EU approval patients for all 3 cohorts but a slightly higher risk profile. In all 3 cohorts (4.5-6 hours, 3-4.5 hours, and ≤3 hours), the median age was 70 years (P = .22 for the 4.5- to 6-hour cohort and P = .27 for the 3- to 4.5-hour cohort), and the median NIHSS score was 2 points lower in the 4.5- to 6-hour and 3- to 4.5-hour cohorts (10 vs 12 points; P < .01) than in the cohort of patients treated within 3 hours of onset. The median time from hospital admission to treatment was 100 minutes for the 4.5- to 6-hour cohort, 80 minutes for the 3- to 4.5-hour cohort, and 60 minutes for the cohort of patients who were treated within 3 hours of stroke onset (P < .01). The percentage of patients with visible recent infarct at baseline imaging was 36% for the 4.5- to 6-hour cohort, 23% for the 3- to 4.5-hour cohort, and 19% for the cohort of patients who were treated within 3 hours of stroke onset (P < .01). Other baseline parameters did not differ significantly between the 4.5- to 6-hour cohort and the cohort of patients who were treated within 3 hours of stroke onset. In the 3- to 4.5-hour cohort, there was a higher percentage of women compared with the cohort of patients who were treated within 3 hours of stroke onset (46% vs 42%; P < .01).

The rates of hemorrhage detected on the postthrombolysis imaging scans (Table 2) were not statistically significantly different among the groups, except for local hemorrhages between patients treated within 3 to 4.5 hours and patients treated within 3 hours, in which the rate of hemorrhage was lower in the patients treated within 3 to 4.5 hours of stroke onset. Figure 2 shows the mRS scores at 3 months. The proportions of patients within each mRS score at 3 months were similar across the cohorts.

The analysis of patients with signs of recent infarct at baseline imaging showed no statistically significant differences in functional independence (mRS score of 0-2) at 3 months between the cohort of patients treated within 4.5 to 6 hours and the cohort of patients treated within 3 hours, for patients with (54% vs 51%; P > .05) or without (69% vs 61%; P > .05) signs of recent infarction. The patients treated within 3 to 4.5 hours had a better outcome than the patients treated within 3 hours, with (56% vs 51%; P < .05) or without signs of recent infarction (65% vs 61%; P < .05).

Table 3 shows the main outcome results in the EU approval criteria cohort. The rates of SICH, the mortality rates, and the functional outcomes did not differ significantly between the cohort of patients treated within 4.5 to 6 hours and the cohort of patients treated within 3 hours. The SICH rates and the mortality rates did not differ between the cohort of patients treated within 3 to 4.5 hours and the cohort of patients treated within 3 hours, but the cohort of patients treated within 3 to 4.5 hours had a higher independence rate than did the cohort of patients treated within 3 hours. In the multivariate analysis, after adjustment for other prognostic factors, there were no statistically significant differences in SICH rates and outcomes at 3 months between the 4.5- to 6-hour and 3- to 4.5-hour cohorts compared with the cohort of patients treated within 3 hours, except for a lower OR for no or minimal disability in the 3- to 4.5-hour cohort. In multivariate analysis, which only adjusted for age and stroke severity, the cohort of patients treated within 4.5 to 6 hours and the cohort of patients treated within 3 hours had similar results, but the cohort of patients treated within 3 to 4.5 hours had poor outcomes at 3 months compared with cohort of patients treated within 3 hours (P < .01).

An unadjusted correlation analysis of the association between time from stroke onset to treatment as continuous variable and mRS score at 3 months was not significant (Pearson correlation, −0.009 [P > .05]; Spearman rank correlation, −0.01 [P > .05]; ordinal logistic regression [P = .09]). After adjustment for age and baseline NIHSS score, time from stroke onset to treatment as a continuous variable was significantly associated with a higher OR for SICH (OR, 1.003 [95% CI, 1.001-1.005]; P < .01, per SITS-MOST; OR, 1.002 [95% CI, 1.00-1.003]; P < .01, per ECASS II; and OR, 1.001 [95% CI, 1.00-1.003] P < .01, per National Institute of Neurological Disorders and Stroke) and for mortality (OR, 1.002 [95% CI, 1.001-1.003]; P < .01) and with a lower OR for no or minimal disability (OR, 0.998 [95% CI, 0.998-0.999]; P < .01) and for functional independence (OR, 0.998 [95% CI, 0.997-0.999]; P < .01) at 3 months. In a subgroup of patients (~5000 patients), additional adjustment for baseline computed tomographic/magnetic resonance angiographic data was provided, and time from stroke onset to treatment as a continuous variable was significantly associated with a lower OR for no or minimal disability (OR, 0.999 [95% CI, 0.997-1.00]; P = .045) and for functional independence (OR, 0.998 [95% CI, 0.997-1.00]; P = .01) at 3 months.

There was no statistically significant difference in any primary cause of death between the cohort of patients treated within 4.5 to 6 hours and the cohort of patients treated within 3 to 4.5 hours, compared with the cohort of patients treated within 3 hours. Intracerebral hemorrhage as primary cause of death was recorded for 1.9% (4 of 212) of patients treated within 4.5 to 6 hours (P = .36), 1.4% (42 of 3032) of patients treated within 3 to 4.5 hours (P = .37), and 1.2% (249 of 21 135) of patients treated within 3 hours. A combination of cerebral infarct and hemorrhage was the primary cause of death for 0.5% (1 of 212) of patients treated within 4.5 to 6 hours (P = .34), 1.4% (42 of 3032) of patients treated within 3 to 4.5 hours (P = .35), and 1.2% (248 of 21 135) of patients treated within 3 hours. A cerebral infarct was the most common cause of death for 6.1% (13 of 212) of patients treated within 4.5 to 6 hours (P = .33), 4.5% (137 of 3032) of patients treated within 3 to 4.5 hours (P = .44), and 4.8% (1016 of 21 135) of patients treated within 3 hours.

Secondary analyses, which included patients with other EU approval violations, showed no statistically significant differences in SICH and 3-month outcomes between patients treated within 4.5 to 6 hours and patients treated within 3 hours in the unadjusted and adjusted analyses. In the unadjusted analysis, patients treated within 3 to 4.5 hours had a significantly higher rate of SICH (2.1% vs 1.7%; P = .02), functional independence (57% vs 55%; P = .008), and no or minimal disability (41% vs 39%; P = .001) compared with patients treated within 3 hours. In the adjusted analysis, patients treated within 3 to 4.5 hours had a statistically significant higher OR for SICH (OR, 1.28 [95% CI, 1.03-1.58]; P = .02) and for mortality at 3 months (OR, 1.12 [95% CI, 1.01-1.25]; P = .03) and a lower OR for functional independence (OR, 0.90 [95% CI, 0.83-0.97]; P = .01) at 3 months compared with patients treated within 3 hours.

To our knowledge, this is the first observational study demonstrating that intravenous alteplase therapy within 4.5 to 6 hours of stroke onset for patients compliant with other EU approval criteria resulted in comparable rates of SICH, mortality, and functional independence to treatment initiated within 3 hours. This observation persisted in the multivariate analysis after adjustment for baseline imbalances. The results were similar in a sensitivity analysis that included other patients not in compliance with EU approval criteria. The updated analysis of the 3- to 4.5-hour cohort for patients in compliance with EU approval criteria confirmed the previous observations^4,9 that the outcomes were comparable to those of patients treated within 3 hours. In the sensitivity analyses, the 3-month outcome was favorable for patients treated within 3 to 4.5 hours compared with patients treated within 3 hours in the unadjusted analysis, but it was less favorable for patients treated within 3 to 4.5 hours compared with patients treated within 3 hours in the adjusted analysis. A trend toward a less favorable outcome in the adjusted analysis was already observed in the previous SITS-ISTR studies^4,9 of patients treated within 3 to 4.5 hours, but this trend now reaches statistical significance, perhaps because of our larger sample size compared with the previous studies.^4,9 It is important to note that a longer duration of time from stroke onset to treatment, as a continuous variable, was significantly associated with higher SICH rates and poor 3-month outcomes after adjustment for age and baseline NIHSS score. This emphasizes the value of early treatment, which is well established in randomized controlled trials,^5,6 and also indicates that our 4.5- to 6-hour cohort is a highly select group of patients. The apparent conflict between these 2 findings is best resolved by considering that, although delayed treatment carries less benefit and higher risks than earlier initiation, clinicians must be apply tighter selection criteria when considering the initiation of treatment beyond 4.5 hours. We did not demonstrate worse outcomes in select groups of patients treated within 4.5 to 6 hours or within 3 to 4.5 hours compared with patients treated within 3 hours.

There were statistically significant differences in some baseline parameters between the 4.5- to 6-hour and 3- to 4.5-hour cohorts compared with the cohort of patient treated within 3 hours. A few important prognostic factors, such as younger age, less severe stroke as measured by baseline NIHSS score, and less frequency of atrial fibrillation, were in favor of the 4.5- to 6-hour and 3- to 4.5-hour cohorts than the cohort of patient treated within 3 hours; however, some other prognostic factors were not in favor of the 4.5- to 6-hour and 3- to 4.5-hour cohorts, such as visible infarct at baseline imaging scans, higher weight, and obviously very long duration of time from stroke onset to treatment. Moreover, dependency (mRS score of 2-5) before current stroke and history of previous stroke earlier than 3 months were more frequently observed in the cohort of patients treated within 3 to 4.5 hours than in the cohort of patients treated within 3 hours. These baseline differences may have balanced each other because the advantages did not uniformly favor one cohort or the other. Finally, these imbalances were taken into account in the multivariate analysis, and we did not find any statistically significant differences among the 3 cohorts with regard to the main outcome parameters for patients who were otherwise in compliance with EU approval criteria. However, the OR for no or minimal disability (mRS score of 0-1) at 3 months in the cohort of patients treated within 3 to 4.5 hours was lower than the OR for the cohort of patients treated within 3 hours. In our previous SITS-ISTR study⁹ with patients in compliance with EU approval criteria, the 3- to 4.5-hour cohort had a higher adjusted OR for SICH and a lower OR for functional independence, which we did not observe in the present primary analysis with a cohort almost double the size of the previous one. This difference in findings between present and previous results could be due to less severe strokes (1 point lower median NIHSS score) in the present 3- to 4.5-hour cohort than in the previous one. However, the similarities between the present and previous results were that the cohort of patients treated within 3 to 4.5 hours had a more favorable unadjusted 3-month outcome than did the cohort of patients treated within 3 hours, although this advantage disappeared in the multivariate analysis, and that the 3- to 4.5-hour cohort even had a significantly lower OR for no or minimal disability at 3 months. The trend is similar for the 4.5- to 6-hour cohort, but the differences were not statistically significant in univariate and adjusted analyses.

We do not have individual patient data in the SITS-ISTR on the reasons for treatment beyond 4.5 hours. One reason for treatment beyond 4.5 hours may be that the clinicians may have decided not to withhold treatment if the patient had been prepared for and informed about thrombolysis within 4.5 hours but the time limit of 4.5 hours had elapsed owing to unanticipated delays. Another reason may be that, by weighing the risk-benefit ratio for the individual patients (young patients with less severe stroke) based on the point estimate of the pooled analysis, clinicians may have decided that a benefit may exist up to 6 hours after stroke onset.^5,6 Another explanation could be that the investigators may have performed a mismatched imaging study and found salvageable tissue, even though 4.5 hours elapsed since symptom onset. This later explanation is supported by the higher frequency of computed tomography and/or magnetic resonance angiography performed for patients treated within 4.5 to 6 hours or within 3 to 4.5 hours compared with patients treated within 3 hours.

One important observation in the present and previous studies^4,9 was the longer duration of time from hospital admission to treatment for patients treated within 4.5 to 6 hours or within 3 to 4.5 hours compared with patients treated within 3 hours. A possible explanation for this delay in treatment for the 4.5- to 6-hour and 3- to 4.5-hour cohorts could be the addition of a new group of patients who could not have previously been treated when there was a strict 3-hour time limit. Other explanations could be that the delay is due to the human factor (ie, knowing that treatment is approved for up to 4.5 hours rather than ≤3 hours after onset) or simply a mathematical association. Clinicians should make the utmost effort to treat patients as early as possible, bearing in mind that the effects of treatment diminish with time.^12,13

In the recently published third International Stroke Trial,¹⁴ 507 patients (33%) were treated in the 4.5- to 6-hour cohort. There were no statistically significant differences in the primary outcomes (alive or independent)¹⁵ at 6 months between the thrombolysis and control groups in the unadjusted (47% vs 42%; P = .13) or adjusted analysis (overall P = .61). “Alive or independent” in the third International Stroke Trial is equal to our functional independence (ie, an mRS score of 0-2). A secondary ordinal analysis provided evidence of a favorable shift in the distribution of Oxford Handicap Scale scores at 6 months with treatment (P < .001). However, the trial¹⁴ was not sufficiently powered to examine the diminishing effect of treatment with time. Our results are not directly comparable to the results of the third International Stroke Trial¹⁴ because we do not have follow-up data at 6 months. An updated systematic review and meta-analysis¹⁶ suggest that the benefit of treatment possibly extends beyond 4.5 hours, perhaps as late as 6 hours for some patients. Our results are consistent with this suggestion for a select group of patients.

One important limitation of our study, which is inherent in any observational study, is the possible nonequivalence of the cohorts, particularly the 4.5- to 6-hour cohort relative to the other 2 cohorts. There is a risk of potential patient selection bias for treatment beyond the 4.5-hour time window (eg, for patients who are younger or have a less severe stroke). The sample size for the 4.5- to 6-hour cohort is rather small and represents only 1% of the total, but our 4.5- to 6-hour cohort, including those not in compliance with EU approval criteria, is comparable in size to the third International Stroke Trial¹⁴ 4.5- to 6-hour cohort (583 vs 507). Although we performed a multivariate analysis to adjust for recorded baseline differences, this may not account for all imbalances.

In conclusion, our observational data suggest that intravenous alteplase given within 4.5 to 6 hours of stroke onset in a highly select group of patients was associated with comparable rates of SICH, mortality, and functional independence to intravenous alteplase given within 3 hours of onset, although a longer duration of time from hospital admission to treatment, as a continuous variable, was significantly associated with higher SICH rates and poor 3-month outcomes. We cannot recommend intravenous thrombolysis beyond 4.5 hours based on these observational data, which require confirmation in controlled studies. The treatment remains safe and effective in the 3- to 4.5-hour time window (compared with the ≤3-hour time window) when used for patients in compliance with EU approval criteria but is slightly less favorable when used for patients not in compliance.

Group Information: A list of the SITS investigators was published in Lancet. 2007;369(9558):275-282.

Accepted for Publication: February 6, 2013.

Corresponding Author: Niaz Ahmed, MD, Department of Neurology, Karolinska University Hospital, SE-171 76 Stockholm, Sweden (niaz.ahmed@karolinska.se).

Published Online: May 20, 2013. doi:10.1001/jamaneurol.2013.406

Author Contributions:Study concept and design: Ahmed.

Acquisition of data: All authors.

Analysis and interpretation of data: Ahmed, Kellert, Mikulik, Tatlisumak, Toni.

Drafting of the manuscript: Ahmed.

Critical revision of the manuscript for important intellectual content: Kellert, Lees, Mikulik, Tatlisumak, Toni.

Statistical analysis: Ahmed.

Obtained funding: Ahmed.

Administrative, technical, and material support: Tatlisumak.

Study supervision: Kellert, Lees, Mikulik, Toni.

Conflict of Interest Disclosures: Dr Ahmed is an employee of SITS International, which received a grant from Boehringer Ingelheim and Ferrer for the SITS-MOST/SITS-ISTR study with alteplase. Dr Lees has received fees and expenses from Boehringer Ingelheim for his role as chairman of the independent data and safety monitoring board of the ECASS III trial with alteplase and for related lectures. He has also received fees from Paion, Forest, and Lundbeck for the Desmoteplase in Acute Ischemic Stroke trials. His institution has received grant assistance toward administrative expenses for coordination of SITS in the United Kingdom. Dr Mikulik has received honoraria payments and travel support from Boehringer Ingelheim. Dr Tatlisumak has significant research contracts with Boehringer Ingelheim, sanofi-aventis, H. Lundbeck A/S, Mitsubishi Pharma, Schering Plough, Concentric Medical, PhotoThera, and BrainsGate. He has received a grant from Boehringer Ingelheim and served on the scientific advisory board of and as consultant for Boehringer Ingelheim, Mitsubishi Pharma, and BrainsGate. Dr Toni received consulting fees from Boehringer Ingelheim and lecture fees from Boehringer Ingelheim, sanofi-aventis, and Pfizer

Funding/Support: The SITS-ISTR is funded by Boehringer Ingelheim, Ferrer, the EU Public Health Executive Authority, a medical training and research fund from the Stockholm County Council and the Karolinska Institutet, the Fighting Stroke project supported by the Swedish Heart-Lung Foundation and Karolinska Institutet, the Friends of Karolinska Institutet in the United States, and Johanniterorden. Dr Mikulik has received research support from the European Regional Development Fund (Project FNUSA-ICRC [grant CZ.1.05/1.1.00/02.0123]).

Additional Information: The Uppsala Clinical Research Centre in Sweden developed, maintained, and upgraded the software for the SITS register in close collaboration with SITS investigators until September 2010. The current SITS registry is developed, maintained, and upgraded by Zitelab (Copenhagen, Denmark) in close collaboration with SITS investigators.

Additional Contributions: We thank all the SITS-ISTR investigators and their centers for their participation. We also thank all the patients who participated in the SITS-ISTR study.