Reperfusion Therapies for Wake-Up Stroke
Patients who present with stroke symptoms on awakening, often referred to as wake-up stroke (WUS), account for ≈25% of acute ischemic strokes.1 In this group, onset time is unknown and usually defined conservatively as time last seen well (LSW) before sleeping.2,3 Onset time is important because benefit from intravenous thrombolysis is time dependent until 4.5 hours after symptoms start.4 Patients with WUS are not treated routinely unless awake and symptom free within standard timescales. However, evidence suggests that clinical and imaging characteristics of patients with WUS do not differ significantly from those with stroke while awake (SWA) presenting within 0- to 3-hour thrombolysis time windows.1,5,6 Imaging findings for many patients with WUS are consistent with an onset immediately before waking.6 As patients with WUS may benefit from reperfusion treatment if promptly delivered, we systematically reviewed studies of patients with WUS treated with intravenous and intra-arterial reperfusion therapies to consider the strength of evidence and make future recommendations.
Search Strategy/Selection Criteria
An unlimited PubMed search (MEDLINE) was performed on April 16, 2013, using the terms: wake* OR awake* OR sleep* OR asleep OR unclear-onset OR unclear onset OR unwitness* OR witness* AND stroke. Further PubMed alerts were received until June 30, 2013. One author (D.B.) reviewed abstracts to identify potentially eligible studies. Because of the small number of publications anticipated, all study designs were included. Articles were included if they described additional analyses using data from previously identified cohorts.
Two authors (D.B. and L.S.) independently extracted data. Discrepancies were resolved by discussion with C.P. as the final adjudicator.
The data extracted were
type(s) of data comparison/numbers of patients,
type(s) of intervention, drugs/doses,
time to treatment (TTT) from wake-up and LSW,
stroke severity (baseline National Institute of Health Stroke Scale [NIHSS]), and
treatment outcomes: disability (modified Rankin Scale [mRS]), Barthel Index, mortality, asymptomatic brain hemorrhage, and symptomatic brain hemorrhage (SBH). Asymptomatic brain hemorrhage and SBH definitions varied (Table I in the online-only Data Supplement).
Synthesis of Results
Because of considerable heterogeneity of reported methods and data, meta-analysis was not feasible and no statistical examination was performed. An organized narrative summary is presented.
The initial search yielded 881 articles, of which 36 were deemed potentially eligible. Eleven were retained for inclusion.7–18 Reasons for exclusion were nontreatment studies, failure to report WUS outcome by treatment group, and patients with WUS within broader unknown onset cohorts.
Characteristics of 12 articles reporting treatment outcomes for 10 WUS cohorts are shown in Table II in the online-only Data Supplement: 3 were case reports/series,7–9 7 were observational stroke registry/database analyses,10–16 2 described the same randomized controlled trial (RCT) containing a cohort of patients with WUS17 with separately published secondary analysis.18 Including subgroup analysis based on additional criteria, there were 5 comparisons of treated versus untreated WUS and 7 of treated WUS versus treated SWA. Numbers of patients with WUS ranged from 2 in case reports to 68 in a registry study.
Types of Treatment and Time Windows
Five cohorts were treated with intravenous thrombolysis alone using doses that reflected international licensing variations. There was 1 RCT of an unlicensed alternative intravenous agent (abciximab).17,18 Patients in 3 cohorts individually received intra-arterial thrombolysis/thrombectomy±intravenous thrombolysis.7,9,16 Two cohorts contained a mixture of patients who had received intravenous thrombolysis or intra-arterial interventions or both according to timing of presentation, radiological criteria, and clinical judgment.12,13
Time criteria for generating study cohorts varied greatly: 1 was <3 hours of waking with symptoms,17,18 2 <6 hours,10,13 1 <12 hours,16 2 <12 hours of LSW rather than waking time,11,14,15 1 <24 hours of LSW,8 and 3 did not define time criteria.7,9,12 Studies were inconsistent when describing TTT. Four did not report time interval data.9,11,14–16 Four specified a wake-up time7,12 and LSW data.7,10,12,17,18 Two reported first found abnormal time; it was unclear whether this was equivalent to wake-up time.8,13 Mean TTT from LSW was typically ≈10 to 12 hours (range, 4–18.8 hours), approximately twice the mean for patients with SWA.
Additional imaging was often used to assist treatment decisions. One study used noncontrast computed tomography (CT) alone for all patients.17,18 Manawadu et al14 reported a sensitivity analysis of noncontrast CT–only patients, which did not change the balance of outcomes between treated and untreated WUS. In 2 cohorts, some14,15 or all16 patients with WUS were selected using CT perfusion techniques. Six cohorts included patients identified by MRI: 2 also used MR perfusion10,13 and 1 used MR angiography.8 One database study used all imaging techniques in different combinations for different patients.12 One case report used digital subtraction angiography.9
Differences in median baseline NIHSS among treated WUS, untreated WUS, and treated SWA reflected small populations and different study protocols. Reported median baseline NIHSS of treated WUS patients ranged from 10 to 14 points.
The outcomes of interest were not consistently reported.
Treated WUS Versus Untreated WUS
In the RCT of abciximab, the proportion of patients with 3-month mRS 0 to 1 was nonsignificantly lower in the treated compared with the placebo WUS group (9% versus 29%; P=0·1).17 There was a similar but less obvious imbalance for mRS 0 to 2, mortality and a primary efficacy measure, which controlled for stroke severity18 (for details, see footnotes of Table 1). Rates of SBH and asymptomatic brain hemorrhage were higher in the treatment group and although this did not reach statistical significance, it led to suspension of recruitment into the WUS cohort after an interim analysis (Table 1).
Unadjusted outcome data for registry studies showed no difference in favorable (mRS, 0–1) or independent (mRS, 0–2) outcomes at 3 months between treated and untreated WUS. After regression analysis to adjust for baseline variables, which could affect outcome (severity, age, and sex), 2 studies reported a statistically significant advantage after treatment for mRS 0 to 2 at 3 months (odds ratio, 5.2 [1.3–20.3] after intravenous treatment; P=0.02)14 and discharge (28% versus 13% after intravenous and intra-arterial treatment; P=0.006).12 This apparent benefit was despite significant increases in total intracranial hemorrhage (13/68 versus 2/54; P=0.004)14 and mortality (7/46 versus 0/34; P=0.02).12 However, no study reported an increase in SBH. In a study where the decision to treat always involved MR perfusion imaging, no SBH was reported but 3-month mRS did not differ between treated and untreated WUS groups.10 A noncontrast CT–only sensitivity analysis for 1 cohort came close to a statistically significant mortality reduction (2/23 treated versus 10/35 untreated; P=0.07) although mRS was no different.14 Overall, there was no obvious distribution of outcomes according to imaging modality.
Treated WUS Versus Treated SWA
In the abciximab RCT the proportion of favorable and independent outcomes was lower in treated WUS than treated SWA but statistical comparison was not reported.18 SBH rates were significantly higher with treated WUS (4/22 versus 18/375; P=0.03; Table 2).
Among registry studies there was no significant difference reported in mRS outcomes, suggesting a similar treatment effect. In a cohort of mixed intravenous and intra-arterial treatment of patients with WUS, Barreto et al12 reported nonstatistically significant improved rates of favorable and independent outcomes at discharge after adjustment for baseline NIHSS (mRS, 0–1: odds ratio, 0·48 [0.18–1.27]; P=0·14 and mRS, 0–2: odds ratio, 0·64 [0.3–1.38]; P=0·62). With comparison to a treated SWA reference group, Manawadu et al15 found no difference for dependency including sensitivity analysis of patients meeting European Cooperative Acute Stroke Study (ECASS) III criteria, but there was a reduction in mortality (1/33 WUS versus 34/197; P=0.03) among these patients with WUS. Bai et al11 presented a combined outcome of mRS 0 to 1 or NIHSS decrease >8 points, but the proportions in each group were equivalent. No significant difference in SBH was shown between treated WUS and treated SWA patients.
One case series measured disability (mRS) and reported that 1 of 4 patients with WUS had a favorable outcome (mRS, 0–2).8 The other 2 case reports described improved clinical characteristics for each of their 2 patients.7,9 Asymptomatic brain hemorrhage occurred in 4 of 8 patients and SBH in 1 of 8; all received treatment based on MRI findings or digital subtraction angiography. No deaths were reported (Table 3).
In this systematic review, we have examined studies reporting outcomes of patients with WUS treated with intravenous and intra-arterial reperfusion therapies. The published literature was limited, data reporting were inconsistent and studies were generally small. However, the reported median baseline NIHSS for treated WUS patients was similar to the score of 11 from pooled clinical trials of intravenous thrombolysis19 and the score of 12 described by the Safe Implementation of Treatments for Stroke register.20
Meta-analysis was not possible because only 1 study was randomized and there was marked variation in settings, criteria, treatments, and outcome measures. It is important to recognize that data from 1 cohort featured more than once.14 No randomized data were found, which described outcomes related to a treatment currently used in clinical practice. Although the abciximab RCT reported an unfavorable trend against treatment of patients with WUS,17,18 this examined reperfusion using a glycoprotein IIb/IIIa antagonist in a small subgroup of patients presenting within 5 hours of stroke symptoms. Therefore, it would not be appropriate to reach a conclusion about reperfusion WUS treatment from this single study. However, the apparent advantage seen after baseline adjustment for 2 WUS studies12,14 must also be interpreted cautiously as the cohorts were small, nonrandomized, and unblinded during protocols, which relied on local clinical judgment for patient selection.
The effect of treatment on disability was inconsistent and depended on whether the reference group was untreated WUS or treated SWA. Excluding results derived from the abciximab trial17,18 or those adjusted for baseline characteristics,12,14 there was no difference in mRS when treated and untreated WUS were compared, but there was also no difference when treated WUS was compared with treated SWA. This probably reflects the small sample sizes limiting detection of any treatment effect and the impact of patient selection by clinical criteria across all groups. A clinical influence is apparent in the variation in absolute numbers of patients with favorable or independent mRS outcomes in studies, ranging from 13% to 30% for mRS 0 to 1 and 30% to 58% for mRS 0 to 2 after any treated WUS and from 3% to 31% for mRS 0 to 1 and 20% to 60% for mRS 0 to 2 in untreated WUS. After treatment of SWA the dependency outcome ranged from 24% to 32% for mRS 0 to 1 and 18% to 51% for mRS 0 to 2. By comparison, 35% of patients allocated recombinant tissue-type plasminogen activator in 10 clinical trials of intravenous thrombolysis had an mRS 0 to 1 and 46% had mRS 0 to 2 versus control results of 29% and 42%, respectively.21 It would not be surprising if WUS outcomes were inferior to these because of a likely shift of TTT distribution toward longer intervals.
Overall, there was no clear effect on mortality, including both an increase12 and a reduction15 after WUS treatment. However, as intravenous thrombolysis has not shown a beneficial effect on short-term mortality in clinical trials,21 it would be surprising if reperfusion treatment reduced WUS mortality.
Comparison of hemorrhage rates was limited by the differing event definitions and reporting time points. Asymptomatic hemorrhage rates ranged from 9% to 19% for treated WUS cohorts. Only the abciximab trial found an excess of symptomatic bleeding during comparison of treated WUS and treated SWA.17,18 However, an estimation of harm from WUS reperfusion treatment is not possible without trials including sufficient numbers of patients, a standard definition of symptomatic hemorrhage, and blinded assessment. When it was clear which imaging modality and treatment had been used, symptomatic hemorrhage rates after intravenous recombinant tissue-type plasminogen activator were 1 of 58 (1.7%) for MRI approaches10,11 and 2 of 68 (2.9%) for CT approaches.14 In this context, it was unclear whether CT perfusion, MRI, and MR perfusion imaging were of additional value, particularly as 1 well-defined noncontrast CT–only subgroup included no symptomatic hemorrhages.14 This suggests that CT selection may be an appropriate imaging modality for future trials of reperfusion therapy in WUS, although current trials are focusing on the possible advantage of MRI estimation of salvageable penumbra.22–24
In addition to variations in patient selection and TTT, the lack of modality-specific data reported by studies with mixed treatment approaches has prevented any conclusion on intravenous and intra-arterial therapy for WUS. As intravenous thrombolysis can be administered more rapidly than intra-arterial therapies, and the latter are being assessed by clinical trials for effectiveness after known recent onset, we recommend intravenous treatment to be the priority for WUS research. Study design should be modified if intra-arterial treatments are proven to have a longer therapeutic time window and greater effectiveness than recombinant tissue-type plasminogen activator or other intravenous agents in development.
There are limited data on the risks and benefits of reperfusion therapies for WUS. Although the only randomized data do not support the safety or benefit of a glycoprotein IIb/IIIa antagonist (abciximab), there is currently insufficient evidence to make recommendations on the use of other treatments or the imaging approach for patient selection. Results from RCTs are required.
Sources of Funding
G.A. Ford is supported by a National Institute for Health Research Senior Investigator Award. The corresponding author had access to all study data and final responsibility for the decision to submit for publication.
G.A. Ford has been paid lecture fees for attending and speaking at workshops held by Boehringer Ingelheim. His institution has received research funding for stroke-related activities from Boehringer Ingelheim and grant assistance toward administrative expenses for coordination of Safe Implementation of Treatments for Stroke in the United Kingdom. G.A. Ford has also received funding from Lundbeck A/S in relation to participation in the steering committee for DIAS 3 and 4. The other authors report no conflicts.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.114.005126/-/DC1.
- Received February 12, 2014.
- Revision received March 28, 2014.
- Accepted April 1, 2014.
- © 2014 American Heart Association, Inc.
- Fink JN,
- Kumar S,
- Horkan C,
- Linfante I,
- Selim MH,
- Caplan LR,
- et al
- Iosif C,
- Oppenheim C,
- Trystram D,
- Domigo V,
- Méder JF
- Barreto AD,
- Martin-Schild S,
- Hallevi H,
- Morales MM,
- Abraham AT,
- Gonzales NR,
- et al
- Kang DW,
- Sohn SI,
- Hong KS,
- Yu KH,
- Hwang YH,
- Han MK,
- et al
- Manawadu D,
- Bodla S,
- Keep J,
- Jarosz J,
- Kalra L
- Manawadu D,
- Bodla S,
- Jarosz J,
- Keep J,
- Kalra L
- Adams HP Jr.,
- Effron MB,
- Torner J,
- Dávalos A,
- Frayne J,
- Teal P,
- et al
- Adams HP Jr.,
- Leira EC,
- Torner JC,
- Barnathan E,
- Padgett L,
- Effron MB,
- et al
- Hacke W,
- Donnan G,
- Fieschi C,
- Kaste M,
- von Kummer R,
- Broderick JP,
- et al
- Ma H,
- Parsons MW,
- Christensen S,
- Campbell BC,
- Churilov L,
- Connelly A,
- et al
- 24.↵MR WITNESS: a study of intravenous thrombolysis with alteplase in MRI-selected patients. http://www.clinicaltrials.gov/show/NCT01282242. Accessed April 21, 2014.