Romano JG, Smith EE, Liang L, Gardener H, Camp S, Shuey L, Cook A, Campo-Bustillo I, Khatri P, Bhatt DL, Fonarow GC, Sacco RL, Schwamm LH. Outcomes in Mild Acute Ischemic Stroke Treated With Intravenous ThrombolysisA Retrospective Analysis of the Get With the Guidelines–Stroke Registry. JAMA Neurol. 2015;72(4):423-431. doi:10.1001/jamaneurol.2014.4354
Mild strokes have been poorly represented in thrombolytic trials and only a few series have reported outcomes after treatment with intravenous (IV) recombinant tissue plasminogen activator (rtPA) after mild stroke.
To report treatment complications and short-term outcomes in patients with mild stroke who have received treatment with IV rtPA.
Design, Setting, and Participants
Retrospective analysis of patients treated in the emergency department of hospitals that use the Get With the Guidelines–Stroke registry, a prospectively collected quality improvement registry used by hospitals across the United States. Patients were those admitted between May 1, 2010, and October 1, 2012, with acute ischemic stroke within 4.5 hours from symptom onset and a baseline National Institutes of Health Stroke Scale score of 5 or less. Univariable and multivariable analyses were performed to identify factors independently associated with discharge outcomes and treatment complications.
Main Outcomes and Measures
Treatment complications included symptomatic intracranial hemorrhage, life-threatening or serious systemic hemorrhage, other serious complications, and undetermined complications. The short-term outcomes analyzed were in-hospital mortality, discharge to home, independent ambulation at discharge, and length of stay.
Among 33 995 patients who arrived within 4.5 hours of symptom onset and were treated with IV rtPA, 7621 (22.4%) had a National Institutes of Health Stroke Scale score of 5 or less and 5910 had complete data for analysis. Treatment complications were infrequent: symptomatic intracranial hemorrhage, 1.8%; life-threatening or serious systemic hemorrhage, 0.2%; other serious complications, 1.8%; and complications of undetermined cause, 2.4%. Mortality was low (1.3%), but at discharge 30.3% could not ambulate independently, 29.4% could not go directly home, and 73.0% had a length of stay of 3 days or longer. Worse short-term outcomes were seen in older patients, African American patients, diabetic patients, and those who arrived by ambulance, after hours, or with a higher National Institutes of Health Stroke Scale score.
Conclusions and Relevance
Many patients with ischemic stroke treated with IV rtPA have a mild stroke. Symptomatic intracranial hemorrhage is infrequent, but approximately 30% of these patients are unable to return directly home or ambulate independently at discharge. Additional studies are needed to identify strategies to improve the outcomes in patients with mild stroke who receive thrombolysis.
Quiz Ref IDCompelling evidence and current guidelines recommend treatment with intravenous (IV) thrombolysis for eligible patients with acute ischemic stroke,1- 3 but recent reports suggest that fewer than 5% of all patients with ischemic stroke receive thrombolytics.4 Mild or rapidly improving stroke symptoms are the most common reasons for lack of treatment in otherwise eligible patients, present in 31% to 43% of patients who arrive within 3 hours of symptom onset.5- 8Quiz Ref IDAlthough thrombolytic administration in patients with mild stroke has increased during the last decade in the United States,9 the efficacy and safety outcomes of thrombolysis in mild stroke remain poorly understood, as only small series are available.
The objectives of this study were to characterize treatment complications and short-term outcomes of patients with mild acute ischemic stroke treated with IV recombinant tissue plasminogen activator (rtPA) within 4.5 hours from symptom onset in the Get With the Guidelines–Stroke (GWTG-Stroke) registry and to analyze complications and outcomes by treatment window (0-3 vs 3-4.5 hours). Secondarily, we aimed to determine the proportion of mild strokes treated with thrombolytics.
The GWTG-Stroke registry, designed as a quality improvement registry and sponsored by the American Heart Association, is used by almost 1700 hospitals across the United States. Trained hospital personnel use an Internet-based tool (Quintiles Outcome) to enter deidentified data into the GWTG-Stroke registry. The Duke Clinical Research Institute serves as the data analysis center. The Duke Clinical Research Institute has institutional review board approval to analyze these data; the University of Miami Institutional Review Board also approved this analysis. This retrospective analysis of previously collected data does not require informed consent as all data are deidentified.
The study population included GWTG-Stroke registry participants with a final diagnosis of acute ischemic stroke admitted between May 1, 2010, and October 1, 2012 (n = 574 342 from 1682 hospitals). We excluded patients from hospitals that had less than 75% completion on medical history variables, which reflects suboptimal documentation (n = 19 288); those who arrived beyond 4.5 hours from symptom onset (n = 407 137); those not treated with IV rtPA (n = 108 096); those with a National Institutes of Health Stroke Scale (NIHSS) score higher than 5 (n = 28 402); those missing an NIHSS score (n = 3176); those who did not arrive through the emergency department or were not discharged from the same hospital, which limits the availability of baseline information and discharge outcomes (n = 1922); and those with a time to treatment that was longer than 4.5 hours or missing (n = 411). A total of 5910 patients from 966 hospitals were included.
The characterization of mild stroke as an NIHSS score between 0 and 5 is based on the commonly accepted practice of using this range to exclude patients from acute treatment trials10- 12 and to define mild stroke in population-based studies.13 The GWTG-Stroke hospitals are instructed to record the first NIHSS score obtained; subsequent changes in the degree of disability are not recorded.
The study population was described with respect to demographic characteristics, insurance, arrival and admission data, medical history, vital signs at presentation, laboratory test results, prior medication use, initial examination findings (GWTG-Stroke collects the baseline NIHSS score and separately collects 4 reported stroke symptoms: weakness or paresis, aphasia, altered consciousness, and other symptoms), in-hospital treatments, acute care (involvement of the neurology service, involvement of a stroke team, or admission to a stroke unit), hospital characteristics and quality measures, treatment complications, and short-term outcomes of interest stratified by treatment window (0-3 vs 3-4.5 hours). Stroke unit care was defined according to the Brain Attack Coalition recommendations.14 Treatment complications were coded in the following manner: (1) symptomatic intracerebral hemorrhage (ICH): brain hemorrhage confirmed by computed tomography within 36 hours of IV rtPA and physician’s notes indicating clinical deterioration due to the hemorrhage; (2) life-threatening or serious systemic hemorrhage: bleeding within 36 hours of IV rtPA requiring transfusion of more than 3 units of blood; (3) other serious complications: complications unexpected or out of proportion to the patient’s expected course, documented as complications of reperfusion therapy such as rapid development of malignant edema, angioedema, or recurrent stroke, and requiring additional medical interventions or prolonged length of stay (LOS); and (4) complications of undetermined cause: worsening stroke symptoms after thrombolysis without definitive evidence to classify the nature of the complication and without death. Short-term outcomes at discharge included discharge to home, ability to walk independently, LOS, and in-hospital mortality. For the outcome of independent ambulation, we restricted the population to those with documented ability to ambulate independently at baseline.
In a separate analysis, we described the proportion of mild strokes among all patients treated with IV rtPA during the same 29-month study period, including all patients treated with IV rtPA 4.5 hours or sooner from symptom onset and with a documented baseline NIHSS score, excluding hospitals with less than 75% completion of medical history variable documentation.
We used χ2 tests to compare categorical variables and Wilcoxon rank sum tests for continuous and ordinal variables. We evaluated the association of patient and hospital variables with short-term outcomes at discharge and treatment complications in a multivariable logistic regression model using a generalized estimating equations approach to account for within-hospital clustering. Covariates in the model include patient baseline characteristics (age, sex, race/ethnicity, insurance, calendar year), medical history (atrial fibrillation or flutter, prosthetic heart valve, previous stroke or transient ischemic attack, coronary artery disease or prior myocardial infarction, carotid stenosis, peripheral vascular disease, hypertension, diabetes mellitus, dyslipidemia, heart failure, and smoking), arrival characteristics (time from symptom onset to arrival, arrival at off hours [between 6 pm and 7 am Monday through Friday; weekends; and holidays], ambulance transport), admission findings (NIHSS score, symptoms of aphasia, hemiparesis, altered consciousness), care unit (admission by neurology service, stroke consultation, stroke unit care), medications prior to admission (anticoagulant, antiplatelet, antihypertensive, cholesterol reducer, and diabetic medications), and hospital characteristics (region, hospital size, stroke volume, IV rtPA volume, teaching status, rural vs urban location). All P values were 2-sided and statistical significance was defined as P < .05. All statistical analyses were performed with SAS version 9.1 software (SAS Institute Inc).
Most variables had missing values in fewer than 5% of cases, except the following categories: symptoms at presentation (18.0%), mode of emergency medical services arrival (9.8%), care unit (14.5%), and medications prior to admission (11.3%-12.4%) except cholesterol reducer. Missing data were imputed to the dominant category for categorical variables and to the median for continuous variables. Missing insurance status for those aged 65 years or older was imputed to Medicare. Missing hospital variables such as teaching status and number of beds (<2%) were excluded.
The baseline patient and hospital characteristics of the 5910 patients from 966 hospitals are described in Table 1. Quiz Ref IDOverall, 98.2% arrived within 3 hours of onset; 78.6% were treated in the 0- to 3-hour window and 21.4% in the 3- to 4.5-hour window. The time from symptom onset to treatment was about 1 hour longer for those treated in the 3- to 4.5-hour window (P < .001). Patients treated in the first 3 hours more commonly were older (P < .001), were men (P = .02), were non-Hispanic white (P = .03), had more atrial fibrillation (P < .001), had more dyslipidemia (P = .003), were more likely to be taking prophylactic medications (anticoagulant, P = .01; antihypertensive, P = .001; antilipidemic, P = .02), and were more likely to arrive by ambulance (P < .001). However, the NIHSS score was similar for those treated in the early and extended windows (P = .31).
Treatment complications and discharge outcomes are detailed in Table 2. Mortality was 1.3%, 1.8% had symptomatic ICH, 0.2% had life-threatening or serious systemic hemorrhage, 1.8% had other serious complications, and 2.4% had complications of undetermined cause. However, 29.4% of patients could not be discharged directly home, 30.3% were not able to ambulate independently, and 73.0% had LOS of 3 days or longer. The outcome of independent ambulation at discharge was similar for the 2 treatment windows (0-3 vs 3-4.5 hours) among those with documented ability to ambulate independently prior to the index stroke only (P = .72) and in a sensitivity analysis including those with missing or undocumented baseline ambulation. There were no differences in short-term outcomes and treatment complications among those treated at 0 to 3 hours compared with those treated at 3 to 4.5 hours, except for a lower rate of treatment complications of undetermined cause in those treated in the 0- to 3-hour window than in the 3- to 4.5-hour window (2.1% vs 3.3%, respectively; P = .01). When outcomes were categorized by NIHSS score (Table 3), worse discharge outcomes were noted with increasing NIHSS score, but symptomatic ICH and other treatment complications were similar across this range of NIHSS scores.
Multivariable-adjusted associations between covariates and select discharge outcomes and treatment complications are provided in Table 4. Only covariates with significant associations are included. Heart failure and altered consciousness on arrival were strongly associated with in-hospital mortality (odds ratio [OR] = 2.21 [95% CI, 1.06-4.61] and OR = 3.53 [95% CI, 1.93-6.46], respectively), although these variables did not occur commonly. Symptomatic intracranial hemorrhage occurred more often in older individuals (OR = 1.35 [95% CI, 1.12-1.63]), in those with a history of carotid stenosis (OR = 2.35 [95% CI, 1.07-5.18]), and when altered consciousness was reported as a presenting symptom (OR = 2.70 [95% CI, 1.55-4.71]); dyslipidemia and care in a stroke unit were associated with a lower risk of symptomatic ICH (OR = 0.62 [95% CI, 0.41-0.93] and OR = 0.57 [95% CI, 0.38-0.86], respectively). Older age, being African American, history of diabetes mellitus, arrival by ambulance, arrival at off hours, worse stroke severity, and hemiparesis were independently associated with inability to be discharged home and/or inability to ambulate (Table 4).
During the study period, GWTG-Stroke hospitals collectively treated 36 552 patients with IV rtPA within 4.5 hours from symptom onset, of whom 33 995 had a documented baseline NIHSS score. Of these 33 995 patients, 7621 (22.4%) had a baseline NIHSS score between 0 and 5. Conversely, among all patients with a baseline NIHSS score between 0 and 5 who arrived at the hospital within 4.5 hours from symptom onset, only 13.5% received IV rtPA.
We have characterized a cohort of patients with mild stroke treated with IV rtPA, of whom one-fifth were treated in the expanded treatment window of 3 to 4.5 hours. To our knowledge, this is the largest report of patients with mild stroke who received thrombolysis. This report adds to the characterization of mild treated strokes as most prior data come from small series. There is a paucity of data from randomized clinical trials on the efficacy and safety of IV rtPA in this population. For example, a total of 108 patients with an NIHSS score of 5 or less received thrombolytics in the National Institute of Neurological Disorders and Stroke and European Cooperative Acute Stroke Study 3 trials combined,15,16 while 304 were treated in the third International Stroke Trial.17 A recent pooled analysis of individual patient data from randomized trials showed better outcome in patients with mild ischemic stroke (NIHSS score 0-4) treated with rtPA (OR = 1.48; 95% CI, 1.07-2.06); however, only 9.9% of enrolled patients had mild stroke.18
Quiz Ref IDThe relatively low complication rates associated with use of IV rtPA in mild stroke, with a symptomatic ICH rate of 1.8% and systemic hemorrhage rate of 0.2%, are an important finding in this study. Treatment complications were similar between the 0- to 3-hour and 3- to 4.5-hour windows except those of undetermined cause, which were more common in the 3- to 4.5-hour window than in the 0- to 3-hour window (3.3% vs 2.1%, respectively; P = .01), and treatment complications were similar across all 6 NIHSS categories studied. The symptomatic ICH rate of 1.8% is much lower than in patients enrolled in clinical trials with a higher NIHSS score, where it has been reported to be 6.4% to 7.9%.1,2,19 While the definition of symptomatic ICH used in GWTG-Stroke is not identical to the National Institute of Neurological Disorders and Stroke criteria,1,20 it is similar and reasonably comparable. In smaller series of mild stroke treated with IV rtPA (27-488 patients), the risk of symptomatic ICH varied from 0% to 5%,10,21- 28 and similar to our results, no increased risk was found in those treated in the extended window.19,24
We identified a higher risk of thrombolytic treatment in older individuals, with an OR of 1.35 (95% CI, 1.12-1.63) for symptomatic ICH for each 10 years of age, consistent with the known association of advancing age and symptomatic ICH.29 A higher risk of symptomatic ICH was also found in those presenting with altered level of consciousness and in patients with a history of carotid stenosis, although it should be noted that 2.7% had a history of carotid disease, 6.6% had altered consciousness at presentation, and there was a high rate of missing presentation symptoms. While altered consciousness is not normally associated with mild stroke, one could envision scenarios in which altered consciousness was present on arrival but improved rapidly prior to the first NIHSS score being obtained.
Care in a stroke unit and history of dyslipidemia were associated with a lower risk of symptomatic ICH. Stroke unit admission was present in only two-thirds of patients, although care in a specialized unit has been clearly shown to reduce complications and improve outcomes and is recommended for all patients who have received thrombolysis.3 However, it should be recognized that patient trajectories across different units in the hospital (emergency department, intensive care unit, stroke unit, general ward, and rehabilitation) are complex and influenced by various patient-related factors as well as hospital resources, and the nomenclature of units may vary across institutions. The association with dyslipidemia deserves further discussion. In GWTG-Stroke, a history of dyslipidemia is based on self-report, prior documentation, or the use of antilipidemic medication. Although better outcomes in acute ischemic stroke have been reported in those with prior statin therapy,30 a number of studies have failed to demonstrate an association between total cholesterol or low-density lipoprotein cholesterol level and symptomatic ICH,31 although contradictory results have been reported for high-density lipoprotein cholesterol and triglycerides levels.32,33 Similar to other studies,32- 34 we did not find an association between symptomatic ICH after thrombolysis and the use of lipid-lowering agents.
Our results also demonstrate that a significant proportion of patients with mild stroke treated with IV rtPA have suboptimal outcomes at discharge. Overall, 30.3% did not achieve independent ambulation by discharge, 29.4% could not return directly home, 73.0% had a prolonged LOS of 3 days or longer, and 1.3% died. With the exception of death, these short-term outcomes were worse with increasing NIHSS score, reflecting greater stroke severity. A small number of patients with an NIHSS score of 0 were treated; this is probably related to the inability of the NIHSS to capture certain cerebellar and brainstem symptoms.35 In addition, as the GWTG-Stroke registry records only 1 point of stroke severity assessment at baseline, it is possible that some patients may have worsened before treatment. However, treatment in the extended window of 3 to 4.5 hours resulted in outcomes similar to those treated earlier. Others have reported similar short-term outcomes.36 Longer-term poor outcomes, defined as a modified Rankin Scale score of 2 to 6 at 3 months, have been reported in 17% to 42%, although these reports come from smaller series.15,21,22,24,25,27 Recent reports suggest that IV rtPA is associated with better delayed outcomes than controls.18,28
We found worse short-term discharge outcomes in older individuals, African American individuals, women, and those with diabetes mellitus, higher baseline NIHSS scores, reported altered level of consciousness on arrival, hemiparesis on arrival, arrival by ambulance, and arrival at off hours. These associations are expected as worse outcomes are expected in older and diabetic patients, while the arrival symptoms and mode of transportation probably reflect larger and more severe strokes.
There are many possible explanations for poor functional outcomes in mild stroke. First, a low NIHSS score does not exclude disabling symptoms37: isolated aphasia, hemianopsia, or monoplegia may result in a low NIHSS score but these symptoms may be disabling. Also, early progression of symptoms and infarct growth may affect 9% to 10% of those with minor stroke and are associated with large-vessel occlusion and hypoperfusion.38- 40 Moreover, early recurrence after rapid improvement has been described41; this may represent vascular reocclusion, described in up to 12% of those treated with IV rtPA,42 failure of collateral flow to sustain penumbral tissue, or even small-vessel occlusion as in the capsular warning syndrome.43 We found subsequent deterioration in 2.4% that was not explained by symptomatic ICH; this was more frequent in the 3- to 4.5-hour treatment window. One may hypothesize that this low rate of worsening may correspond to a lower frequency of large-vessel occlusion. However, this cannot be ascertained from our data.
Although we have characterized short-term outcomes in mild stroke, the long-term outcomes have not yet been well studied in a large controlled cohort. As most individuals who go to rehabilitation achieve their highest level of independence at 12 weeks,44 3-month outcome is a better measure of disability. As noted, smaller series have reported on the modified Rankin Scale, the most common outcome measure used in stroke studies, but this scale has limitations such as high interrater variability45 and unequal disability weight for each level of the scale.46 Also, it does not interrogate quality of life and activities of daily living and therefore may not capture the full range of disability associated with mild stroke.47,48 We did not compare the outcomes of patients with mild stroke treated with rtPA with the outcomes of those who were not treated, owing to the important selection bias introduced by treatment decisions for which we cannot adequately adjust because of measured and unmeasured residual confounding. This will require randomized prospective trials such as the Potential for rtPA to Improve Stroke With Mild Symptoms study (PRISMS; clinicaltrials.gov identifier NCT02072226). The PRISMS trial and the Mild and Rapidly Improving Stroke Study (MaRISS; clinicaltrials.gov identifier NCT02072681), an observational study, will provide detailed multidimensional measures of the dynamic nature of mild stroke and the longer-term outcomes that will discern the impact and full range of disability associated with mild stroke.
Limitations of this study include those associated with registry-derived data. Hospital participation in the GWTG-Stroke registry is voluntary, and it is possible that our population does not fully reflect the experience across the United States. Indeed, a larger proportion of patients in our sample was treated at teaching hospitals compared with a nationwide inpatient sample.49 However, GWTG-Stroke is used by 1 in 3 acute care hospitals, and with more than 3 million patient entries, it encompasses 1 in 4 strokes in the country. Moreover, prior studies with linkage to national Medicare data suggest that data from the GWTG-Stroke registry are representative of the national Medicare stroke population.50 Another limitation is that we have only 1 fixed point of stroke severity assessment, and some patients with mild stroke may have had more or less severe deficits prior to the baseline recorded NIHSS score and rtPA administration. Residual measured or unmeasured confounding may have influenced the findings.
This study also has important strengths that need to be underscored. To our knowledge, this is the largest study to date of mild stroke treated with IV rtPA. It carefully delineates the factors associated with poor discharge outcomes and expands prior reports8 by including those treated in the expanded window of 4.5 hours.51 The period of data collection (2010-2012) and the large number of hospitals provide a current representation of the characteristics and outcomes of this population in the United States.
Quiz Ref IDWe have described the treatment complications and short-term outcomes and identified baseline predictors of outcomes in mild stroke treated with thrombolytics. The efficacy of thrombolytics in this population cannot be discerned from our study, for which randomized studies are needed. However, given the significant proportion of treated mild stroke and real-world safety and outcomes observed, our results provide reassurance about the safety of IV rtPA in patients with low NIHSS scores.
Corresponding Author: Jose G. Romano, MD, Department of Neurology, University of Miami Miller School of Medicine, 1120 NW 14th St, Ste 1357, Miami, FL 33136 (firstname.lastname@example.org).
Accepted for Publication: November 25, 2014.
Published Online: February 2, 2015. doi:10.1001/jamaneurol.2014.4354.
Author Contributions: Drs Romano and Liang had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Romano, Smith, Gardener, Camp, Shuey, Campo-Bustillo, Sacco, Schwamm.
Acquisition, analysis, or interpretation of data: Romano, Smith, Liang, Gardener, Camp, Shuey, Cook, Khatri, Bhatt, Fonarow, Schwamm.
Drafting of the manuscript: Romano, Smith, Khatri, Bhatt, Fonarow.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Liang, Gardener.
Obtained funding: Romano.
Administrative, technical, or material support: Romano, Smith, Camp, Shuey, Cook, Campo-Bustillo, Fonarow, Sacco.
Study supervision: Romano, Shuey, Cook, Khatri, Schwamm.
Conflict of Interest Disclosures: Dr Romano reported receiving research salary support from Genentech and serving on the scientific advisory board for Vycor/NovaVision. Dr Smith reported serving on the data safety monitoring board of an alteplase trial funded by the National Institute of Neurological Disorders and Stroke. Dr Gardener reported receiving research salary support from Genentech. Dr Campo-Bustillo reported receiving research salary support from Genentech. Dr Khatri reported receiving research salary support from Genentech, Penumbra, and Biogen; honorarium from UpToDate; and book royalties from Taylor and Francis. Dr Bhatt reported serving on the advisory boards for Elsevier PracticeUpdate Cardiology, Medscape Cardiology, and Regado Biosciences; serving on the board of directors of the Boston VA Research Institute and the Society of Cardiovascular Patient Care; serving as chair on the American Heart Association Get With the Guidelines steering committee; serving on data monitoring committees for Duke Clinical Research Institute, Harvard Clinical Research Institute, Mayo Clinic, Population Health Research Institute; receiving honoraria from the American College of Cardiology (editor, Clinical Trials, Cardiosource), Belvoir Publications (editor in chief, Harvard Heart Letter), Duke Clinical Research Institute (clinical trial steering committees), Harvard Clinical Research Institute (clinical trial steering committee), HMP Communications (editor in chief, Journal of Invasive Cardiology), Population Health Research Institute (clinical trial steering committee), Slack Publications (chief medical editor, Cardiology Today’s Intervention), and WebMD (continuing medical education steering committees); serving as deputy editor of Clinical Cardiology and pharmacology section editor of the Journal of the American College of Cardiology; receiving research grants from Amarin, AstraZeneca, Bristol-Myers Squibb, Eisai, Ethicon, Medtronic, Roche, Sanofi-Aventis, and The Medicines Company; and performing unfunded research for FlowCo, PLx Pharma, and Takeda. Dr Fonarow is an employee of the University of California, which holds a patent on a retriever device for stroke. Dr Sacco reported serving as a consultant to Boehringer-Ingelheim. Dr Schwamm reported serving as the principal investigator of a National Institute of Neurological Disorders and Stroke–funded study for which Genentech provides alteplase and supplementary site payments and serving as a member of the international steering committee of desmoteplase trials for Lundbeck. No other disclosures were reported.
Funding/Support: This work was supported by a grant from Genentech to the University of Miami.
Role of the Funder/Sponsor: Genentech had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Previous Presentation: This work was presented in part as a poster at the International Stroke Conference 2014; February 12, 2014; San Diego, California.