Guideline Acute medical and surgical management- …...Acute medical and surgical management- Clinical Guidelines for Stroke Management (2017) Contact Language en Start Date 04.09.2016

Acute medical and surgical management- ClinicalGuidelines for Stroke Management (2017)

Main editor

NSF Stroke guidelines content development working parties

Publishing Info

v0.10 published on 26.09.2016

National Stroke Foundation

1 of 157

Acute medical and surgical management- ClinicalGuidelines for Stroke Management (2017)

Contact

Language

en

Start Date

04.09.2016

Last Edit

22.09.2016

Disclaimer

Acute medical and surgical management- Clinical Guidelines for Stroke Management (2017) - National Stroke Foundation

2 of 157

Sections

1 - Acute medical and surgical management - overview........................................................................................................................................................................ 12

2 - Stroke unit care................................................................................................................................................................................................................................................ 13

3 - Palliative care.................................................................................................................................................................................................................................................... 34

4 - Reperfusion therapy ...................................................................................................................................................................................................................................... 40

4.1 - Thrombolysis....................................................................................................................................................................................................................................... 40

4.2 - Neurointervention ............................................................................................................................................................................................................................ 54

5 - Antithrombotic therapy ............................................................................................................................................................................................................................... 67

6 - Acute blood pressure lowering therapy................................................................................................................................................................................................. 77

7 - Surgery for ischaemic stroke and management of cerebral oedema ......................................................................................................................................... 86

8 - Intracerebral haemorrhage (ICH) management................................................................................................................................................................................. 95

8.1 - Medical interventions...................................................................................................................................................................................................................... 95

8.2 - Surgical interventions................................................................................................................................................................................................................... 106

9 - Oxygen therapy ............................................................................................................................................................................................................................................ 115

10 - Neuroprotection........................................................................................................................................................................................................................................ 121

11 - Glycaemic therapy .................................................................................................................................................................................................................................... 122

12 - Pyrexia management ............................................................................................................................................................................................................................... 141

13 - Dysphagia ..................................................................................................................................................................................................................................................... 147


3 of 157

Summary of recommendations

2 - Stroke unit care

Strong Recommendation

All people with stroke should be admitted to hospital and be treated in a stroke unit with an interdisciplinary team. (SUTC 2013 [1])

Practice Statement

1. All people with stroke should be admitted directly to a stroke unit (preferably within three hours of stroke onset).

2. If people with suspected stroke present to non-stroke unit hospitals, transfer protocols should be developed and used to guide urgent

transfers to the nearest stroke unit hospital.

3. Where transfer is not feasible, sSmaller isolated hospitals should managerconsider stroke services in a manner that adheres as closely

as possible to the criteria for stroke unit care. Where possible, patients should receive care in geographically discrete units.If people with

suspected stroke present to non-stroke unit hospitals, transfer protocols should be developed and used to guide urgent transfers to the

nearest stroke unit hospital.


All acute stroke services should use standarised protocols to manage physiological status and prevent complications. (Middleton et al

2011 [97]).

3 - Palliative care


Stroke patients and their families/carers should have access to specialist palliative care teams as needed and receive care consistent with

the principles and philosophies of palliative care.

Practice Statement

1. An accurate assessment of prognosis or imminent death should be made for patients with severe stroke or those who are

deteriorating.

2. A pathway for stroke palliative care can be used to support stroke patients and their families/carers and improve care for people dying

after stroke.


4 of 157

4.1 - Thrombolysis


• Intravenous alteplase (dose of 0.9 mg/kg , maximum o f 9 0 mg) in acute ischaemic stroke should be administered in patients with

potentially disabling ischaemic stroke meeting specific eligibility criteria (Wardlaw 2014 [20]; Emberson et al. 2014 [21]).

• Thrombolysis should commence as early as possible (within the first few hours) after symptom onset but may be used up to 4.5

hours after onset (Anderson et al. 2016 [23]).

Practice Statement

Thrombolysis should be undertaken in a setting with appropriate infrastructure, facilities and network support (e.g. viatelemedicine) including:

• access to an interdisciplinary acute care team with expert knowledge of stroke management who are trained in delivery of

thrombolysis and monitoring of patients receiving thrombolytic therapy

• a streamlined acute stroke assessment workflow (including ambulance prenotification, code stroke team response and direct

transport from triage to CT scan) to minimise treatment delays and protocols available to guide medical, nursing and allied health

acute phase management

• immediate access to imaging facilities and staff trained to interpret images

• routine data collected in a central register to allow monitoring, benchmarking and improvements of patient outcomes over time

for those treated with reperfusion

• Where possible the patient and caregivers should be involved in the decision to give thrombolysis. However, as an emergency

therapy, formal consent for thrombolysis is not required.

4.2 - Neurointervention


Endovascular thrombectomy commenced within six hours of stroke onset should be undertaken in patients with ischaemic stroke causedby a large vessel occlusion in the internal carotid artery, proximal cerebral artery (M1 segment), or in tandem occlusion of both thecervical carotid and intracranial arteries.


Endovascular thrombectomy should commence as early as possible. If eligible, patients should receive intravenous alteplase whileconcurrently arranging endovascular thrombectomy with neither treatment delaying the other.


5 of 157


Endovascular thrombectomy shouldmay be performedconsidered in selected patients with occlusion of the basilar artery (given the poornatural history).

Needs rationale and any practical considerations

Practice Statement

Endovascular thrombectomy may be considered in the following situations based on individual patient and advanced imaging factors:

• commencement of procedure surgery beyond 6within 12 hours

• occlusion in more distal middle cerebral artery (M2 segment)

• Endovascular thrombectomy should be performed by an experienced neurointerventionist with recognised training in the

procedure (Conjoint Committee for Recognition of Training in Interventional Neuroradiology).

Add Rationale if needed

5 - Antithrombotic therapy


Patients with ischaemic stroke who are not receiving reperfusion therapy should receive antiplatelet therapy as soon as possible once

brain imaging has excluded haemorrhage.

Please review and edit Key info and Rationale as necessary, add practical advice (if any)


Acute antiplatelet therapy should be deferred for 24 hours if alteplase has been given.


Strong Recommendation AGAINST

Routine use of early anticoagulation in unselected patients ( i . e . without c a r d i o e m b o l i s m ( e . g . AF) following TIA/stroke is not

recommended.



6 of 157

Weak Recommendation

Aspirin plus clopidogrel may should be used in the short term (first three weeks) in high risk patients with after minor ischaemic stroke or

TIA to prevent stroke recurrence.


6 - Acute blood pressure lowering therapy

Practice Statement

• All acute stroke patients should have their blood pressure closely monitored in the first 48 hours after stroke onset.

• Patients with acute ischaemic stroke eligible for treatment with intravenous thrombolysis should have their blood pressure

reduced to below 185/110 mmHg before treatment and in the first 24 hours after treatment.

• Patients with acute ischaemic stroke with blood preesure >220/120/mmHg should have their blood pressure cautiously reduced

(e.g. by no more than 20%) over the first 24 hours.

Weak Recommendation AGAINST

Aggressive blood pressure lowering in the acute phase of care to a target SBP of <140mmHg is not recommended (Bath and Krishnan

2014 [57]; Qureshi et al. 2016 [58]; Lee et al. 2015 [56]).

Weak Recommendation

In people with acute primary intracerebral haemorrhage blood pressure may be cautiously reduced with a target SBP of, but not below,

140mmHg (Bath and Krishnan 2014 [57]; Qureshi et al. 2016 [58]).

Weak Recommendation

Pre-existing antihypertensives may be withheld until patients are neurologically stable and treatment can be given safely (Bath and

Krishnan 2014 [57]).

7 - Surgery for ischaemic stroke and management of cerebral oedema


Selected patients aged 60 years and under with malignant middle cerebral artery territory infarction should undergo urgent

neurosurgical assessment for consideration of decompressive hemicraniectomy. When undertaken h e m i c r a n i e c t o m y surgery

should i d e a l l y be performed within 48 hours of symptom onset. (Cruz-Flores et al 2012 [61])

Needs rationale and practical advice (if any)


7 of 157

Weak Recommendation

Decompressive hemicraniectomy may be considered in highly selected patients over the age of 60 years, after careful consideration of

the premorbid functional status , and c o n s i d e r a t i o n o f w h e t h e r i t i s i n t h e patient p r e f e r e n c e s ' s b e s t l o n g - t e r m i n t e

r e s t s. (Back et al 2015 [59]; Jüttler et al. (2014) [60]; Cruz-Flores et al 2012 [61])

Needs rationale and practical advice


Corticosteroids are not recommended for management of patients with brain oedema and raised intracranial pressure. (Sandercock et al

2011 [62])

Needs rationale and practical advice. Meeting doc has as weak rec against so check

Practice Statement

Osmotherapy and hyperventilation can be trialled while a neurosurgical consultation is undertaken.

8.1 - Medical interventions

Weak Recommendation

1. Prothrombin complex concentrate should be used in preference to standard fresh frozen plasma to reverse coagulopathy in patients

with warfarin-related intracerebral haemorrhage.

2. Vitamin K should be used in addition to prothrombin complex to reverse warfarin but is insufficient as a sole treatment.

Please complete rationale and any practical advice


Patients with intracerebral hemorrhage related to DOACs should receive a specific reversal agent where available.

Add GRADE, rational, practical adviceLiterature search didn't identify direct evidence for this recommendations. Please inform Cindy if good evidence is available otherwise suggestchanging it to weak recommendation or practice statement.


8 of 157


In patients with intracerebral haemorrhage previously receiving antiplatelet therapy, we recommend that platelet transfusion should be

avoided.

Please complete Key info, Rationale and any practical advice


Aggressive blood pressure lowering in intracerebral hemorrhage to a target SBP of <140mmHg is not recommended.

Recommendation was copied from section Acute blood pressure lowering. Please review and adjust as necessary

Weak Recommendation

In people with acute primary intracerebral haemorrhage blood pressure may be cautiously reduced with a target SBP of , b u t n o t b e l o

w , 140mmHg.


8.2 - Surgical interventions

Practice Statement

In patients with large (>3cm) cerebellar haemorrhage, decompressive surgery may be offered to selected patients. For other

infratentorial haemorrhages (<3cm cerebellar, brainstem) the value of surgical intervention is unclear.

S u r g i c a l e v a c u a t i o n m a y b e u n d e r t a k e n f o r c e r e b e l l a r h e m i s p h e r e h a e m a t o m a s > 3 c m d i a m e t e r i n s e l e c t

e d p a t i e n t s .


Routine surgical evacuation of supratentorial intracerebral haemorrhage (lobar, basal ganglia and/or thalamic locations) is not

recommended outside the context of research .

Patients with lobar haematomas may be considered for surgery.


Use of intraventricular thrombolysis via external ventricular drain catheter for the treatment of intraventricular haemorrhage is not

recommended outside the context of research.

Please review and edit Key info, Rationale, and any Practical advice


9 of 157

9 - Oxygen therapy


Patients who are hypoxic (i.e. <95% oxygen saturation) should be given supplemental oxygen.

PT: there is no direct evidence for this and it was a practice statement in 2010 guideline - suggest changing this to practice statement and completethe Rationale as needed.


The routine use of supplemental oxygen is not recommended in acute stroke patients who are not hypoxic.

PT: Some differing of opinion on the strength. As there is no clear harm, potential resource implications would make it weak (against) rather thanstrong. Please review and edit key info and rationale accordingly.


Hyperbaric oxygen therapy in acute ischaemic stroke patients is not recommended.


10 - Neuroprotection

Practice Statement

Putative neuroprotective agents, including hypothermic cooling, should only be used in a randomised controlled trial.

Practice Statement

Patients with acute ischaemic stroke who were receiving statins prior to admission can continue statin treatment.

11 - Glycaemic therapy


All patients should have their blood glucose level monitored and appropriate glycaemic therapy instituted to treat hyperglycaemia(glucose levels greater than 10mmols/L) regardless of the patients diabetic status. (Bellolio et al 2014 [85]; Ntaios et al 2014 [86];Middleton et al 2011 [89] and Drury et al 2014 [98])


10 of 157


An early intensive approach to the maintenance of tight glycaemic control (between 4.0-7.5mmol/L) is not recommended. (Bellolio et al2014 [85]; Ntaios et al 2014 [86];

12 - Pyrexia management


All patients should have their temperature monitored at least four times a day for 72 hours. (Middleton et al 2011 [97])

Weak Recommendation

Antipyretic therapy, comprising paracetamol, may be used where fever > 37.5 o C. (den Hertog et al 2009 [102]; Middleton et al 2011 [97])

Practice Statement

Antipyretics can be given orally or via a nasogastric tube or suppository (for those with dysphagia)


11 of 157

1 - Acute medical and surgical management - overview

This chapter covers medical and surgical management in the acute phase of care. Importantly, several other critical components of very earlyassessment (including screening) and management should be routinely provided in addition to those discussed in this chapter. These includeDysphagia, Nutrition and hydration, Incontinence and Deep venous thrombosis (DVT) or pulmonary embolism (PE). Furthermore,rehabilitation should commence in the acute phase (Timing of rehabilitation).


12 of 157

2 - Stroke unit care

The organisation of hospital services to provide stroke unit care is the single most important recommendation for improving strokemanagement. While numbers of stroke units and stroke unit beds have increased between 2010 and 2016, the percentage of patientsreceiving stroke unit care has not increased (Stroke Foundation 2014 ). Therefore stroke unit care should be the highest priority for cliniciansand administrators to consider.

Models of stroke unit care described in the literature include:

• acute stroke ward: acute unit in a discrete ward (usually discharged within seven days)• comprehensive stroke unit care: combined acute and rehabilitation unit in a discrete ward• stroke rehabilitation unit: a discrete rehabilitation unit for stroke patients who are transferred from acute care 1–2 weeks post stroke• mixed rehabilitation ward: rehabilitation provided on a ward managing a general caseload.

The evidence for the benefits of stroke unit care is clearest for units that can provide several weeks of rehabilitation on a comprehensive

stroke unit or stroke rehabilitation unit.5, 38 Different models of rehabilitation produce slightly different results (Table 2). Services that canprovide combined or highly integrated acute and rehabilitation care appear to deliver the best outcomes.

In Australia, most stroke units have a primary focus on acute care and early aspects of rehabilitation, with varying degrees of intensity andfollow-up. There are 68 stroke units managing acute stroke patients but only eight stroke rehabilitation units as reported in the NationalStroke Audits in 2008 and 2009.

The stroke units that have been shown to deliver highly effective stroke care share a number of characteristics including:

• location in a geographically discrete unit• comprehensive assessments• a coordinated multidisciplinary team• early mobilisation and avoidance of bed-rest• staff with a special interest in the management of stroke, and access to ongoing professional education and training• clear communication, with regular team meetings to discuss management (including discharge planning) and other meetings asneeded (e.g. family conferences)• active encouragement of stroke survivors and their carers/families to be involved in the rehabilitation process.

All hospital services should clearly review the existing stroke services in light of the recommendations below. For hospitals without existingstroke units the Stroke Foundation Acute Stroke Services Framework provides details of the minimum standards for acute stroke unit care:the recommended infrastructure, processes, workforce and monitoring which can be used to plan for stroke service improvement. Forhospitals with existing stroke units, consideration should be given to reviewing the percentage of stroke patients actually admitted to thestroke unit to determine if there is adequate capacity (i.e. bed numbers). Clear protocols for bed allocation are needed for all stroke unithospitals.


All people with stroke should be admitted to hospital and be treated in a stroke unit with an interdisciplinary team. (SUTC 2013 [1])

Practical Info

Link to the Acute services framework for defintion of stroke unit [8]

Key Info

Benefits and harms

There is substantial evidence of benefit from organised inpatient stroke unit care of stroke patients. Care must be delivered on in theone area/ward as there is little benefits for mobile stroke teams. There is no evidence of harm from admission of stroke patient tostroke unit.

Substantial net benefits of the recommended alternative


13 of 157

Rationale

Stroke patients who receive organised inpatient care in a stroke unit are more likely to be alive, independent, and living at home one yearafter the stroke based on the 2013 Cochrane review based on 28 trials and 5855 patients. Stroke unit care must comprise as least fourminimum criteria as outlined in the Acute Stroke Services Framework. That care delivered on the one ward by an interdisciplinary teamwho meet at least once a week to plan patient care and who also have professional development specific to stroke.

Quality of evidence

The overall quality of evidence is high based on Cochrane analysis (SUTC 2013 [1]).

High

Preference and values

There appears to be no significant impact of patient preference and values on provision of organised inpatient stroke unit care.

No substantial variability expected

Resources and other considerations

Economic considerationsOrganised inpatient stroke unit care requires a multidisciplinary stroke team lead by stroke physicians and availability of cerebral andcerebrovascular imaging as well as relevant laboratory services.

To date there is only one study investigating economic implications of stroke unit care in Australia (Moodie et al 2006 [5]). It foundstroke unit care to be cost-effective for the delivery of specialist medical services, and for prevention of severe complications. Thefindings may be subject to some bias because of the study's prospective cohort design. International literature found stroke unit careto be either cost-saving or cost-effective over the lifetime (Hunter et al 2013 [6]; Te Ao et al 2012 [7]).

Implementation considerationsThe Australian National Acute Stroke Services Framework clearly defines the services, infrastructure and staff found in a stroke unit(Stroke Foundation 2015 [8]).

No important issues with the recommended alternative

Clinical Question/ PICO

Population: Adults with stroke

Intervention: Organised stroke unit care

Comparator: Alternative services (less organised care)

Summary

A Cochrane review conducted by the Stroke Unit Trialists' Collaboration (2013) [1] compared organised stroke unit care toalternative services. The review included 28 RCTs with 5855 participants. Organised stroke unit care was defined as focused carefor stroke patients by a multidisciplinary team specialising in stroke management. This included:

• Stroke wards where care was given in a discrete ward caring exclusively for stroke patients.• Mixed rehabilitation wards with multidisciplinary teams and specialist nursing staff in a ward that does not careexclusively for stroke patients.• Mobile stroke teams that provide care in a variety of settings

Overall, organised stroke unit care significantly reduced mortality compared to conventional care (OR 0.76, 95% CI 0.66 to 0.88),and significantly reduced the odds of death or institutionalisation and death or dependency. Comparisons between different kindsof stroke wards generally did not provide strong evidence for particular forms of stroke unit organisation.

Sun et al (2013) [2] also carried out a systematic review and meta-analysis comparing acute stroke unit care to conventional carein general medical wards. The 8 trials included in the main analysis were a subset of those included in the Cochrane review for thecomprehensive stroke ward subgroup, excluding unpublished data and trials with short observation periods. Analysis of these 8


14 of 157

trials revealed a borderline significant effect on mortality (OR 0.84, 95% CI 0.70 to 1.00). This is a weaker finding than was seen inthe Cochrane review. It may be explained by the inclusion of newer and unpublished studies in the 2013 Cochrane review.However, Sun et al. also mention an error in an earlier version of the Cochrane review that appears to be uncorrected in the 2013review, where the number of deaths in the control group was incorrectly recorded for one trial. This error does raise somequestions about the effect reported in the 2013 Cochrane review, suggesting the true effect may be slightly weaker.

Chan et al (2013) [3] conducted a systematic review of comparisons between different forms of stroke unit care, i.e. comparisonsacute stroke units, rehabilitation units and comprehensive units which provide both acute care and rehabilitation. There were norandomised controlled trials that directly compared comprehensive stroke units to other forms of stroke unit, so the reviewincluded a meta-analysis which used indirect comparisons, a cross-sectional comparison and a 'before-and-after' study. Thereview found that comprehensive stroke units were associated with decreased death and dependency and shorter length of stay.However, the indirect nature of the evidence means that there is substantial uncertainty about these benefits.

OutcomeTimeframe

Study results andmeasurements

Absolute effect estimates

Alternative services(less organised care)

Organised stroke unitcare

Certainty ineffect

estimates(Quality ofevidence)

Summary

Death by theend of scheduled

follow-upUp to 12 months

7 Critical

Odds Ratio 0.76(CI 95% 0.66 - 0.88)

Based on data from 5,855patients in 28 studies.

(Randomized controlled)Follow up 6 weeks to 12

months, median 12months

221per 1000

177per 1000

Difference: 44 fewer per 1000( CI 95% 63 fewer - 21 fewer )

ModerateSensitivity analysis

based only onthose trials with a

low risk of bias:Stroke unit carewas associated

with a statisticallynon-significant

reduction in theodds of death (OR0.82, 95% CI 0.64to 1.05; P = 0.12)

The people receivingorganised inpatient

(stroke unit) care weremore likely to survive

than those receiving lessorganised care.

Death orinstitutional

care by the endof scheduled

follow-upUp to 12 months

7 Critical

Odds Ratio 0.76(CI 95% 0.67 - 0.86)


(Randomized controlled)Follow up Median follow-

up of 1 year

406per 1000

342per 1000


HighThe sensitivity

analysis based onthose trials that

used anunequivocally

blindedassessment

suggested thatsuch bias has not

seriouslyinfluenced the

results.


(stroke unit) care weremore likely to survive and

return home than thosereceiving less organised

care.

Death ordependency by

the end ofscheduledfollow-up

Odds Ratio 0.8(CI 95% 0.67 - 0.97)


(Randomized controlled)

611per 1000

557per 1000


HighThe main

methodologicaldifficulty when

using dependencyas an outcome was



regain independence


15 of 157

up to 12 months

7 Critical

Follow up Median of 1year

the degree ofblinding at finalassessment andthe potential for

bias if the assessorwas aware of the

treatmentallocation. The

results wereunchanged (OR

0.74, 95% CI 0.61to 0.90; P = 0.002)when restricted to

those trials inwhich an

unequivocallyblinded final

assessment for allparticipants was

undertaken

than those receiving lessorganised care.

Death ordependency at

5-year follow-upFive year follow up

4 Important

Odds Ratio 0.54(CI 95% 0.22 - 1.34)

Based on data from 535patients in 2 studies.

(Randomized controlled)Follow up 5 years

859per 1000

767per 1000

Difference: 92 fewer per 1000( CI 95% 286 fewer - 32 more )

ModerateThere is sustained

benefit amongstroke unit

patients.

There is probablysustained benefit among

stroke unit patients.

Death ordependency at10-year follow-

up10 years

4 Important

Odds Ratio 0.7(CI 95% 0.27 - 1.8)


(Randomized controlled)Follow up 10 years

900per 1000

863per 1000


LowThe summary

results continue tofavour stroke unit

care but withincreased

heterogeneity andloss of statistical

significance for theoutcomes of death

or dependency

There is probablysustained benefit among

stroke unit patients.

Length of stay ina hospital orinstitution or

bothUntil discharge

4 Important

Based on data from:4,115 patients in 18

studies. (Randomizedcontrolled)

Days (n/a) Days (n/a)

Difference: SMD 0.15 fewer( CI 95% 0.32 fewer - 0.02 more )

Very LowThe calculation ofa summary resultfor length of stay

was subject tomajor

methodologicallimitations.

More research isrequired in this area.

Details about studies used and certainty down- and upgrading


16 of 157

Death by the end ofscheduled follow-

up

Intervention reference:Systematic reviewBaseline/comparatorreference: Control arm ofreference used forintervention

Risk of bias: No serious Inadequate/lack of blinding of outcome assessors, resulting inpotential for detection bias ;Inconsistency: No serious ;Indirectness: No seriousImprecision: No seriousPublication bias: No seriousStep 3: Upgrade factors- only if relevant: ~LARGE_EFFECT

Death orinstitutional care by

the end ofscheduled follow-

up


Risk of bias: No serious Inadequate/lack of blinding of outcome assessors, resulting inpotential for detection bias. The sensitivity analysis based on those trials that used anunequivocally blinded assessment suggested that such bias has not seriously influenced theresults. ;Inconsistency: No seriousIndirectness: No seriousImprecision: No seriousPublication bias: No serious

Death ordependency by the

end of scheduledfollow-up


Risk of bias: No serious Inadequate/lack of blinding of outcome assessors, resulting inpotential for detection bias. The sensitivity analysis based on those trials that used anunequivocally blinded assessment suggested that such bias has not seriously influenced theresults. ;Inconsistency: No seriousIndirectness: No seriousImprecision: No seriousPublication bias: No serious


5-year follow-up

Intervention reference:Systematic review [1] withincluded studies:Nottingham 1996,Trondheim 1991,Baseline/comparatorreference: Control arm ofreference used forintervention

Risk of bias: No serious There is unclear bias regarding concealment, blinding of outcomeassesment and incomplete outcome data in the two trials included. ;Inconsistency: Serious The magnitude of statistical heterogeneity was high, with I^2: 72%. ;Indirectness: No seriousImprecision: No seriousPublication bias: No serious


10-year follow-up

Intervention reference:Systematic review [1] withincluded studies:Nottingham 1996,Trondheim 1991,Baseline/comparatorreference: Control arm ofreference used forintervention

Risk of bias: No serious There is unclear bias regarding concealment, blinding of outcomeassesment and incomplete outcome data in the two trials included. ;Inconsistency: Serious The magnitude of statistical heterogeneity was substantial, with I^2:62%. ;Indirectness: No seriousImprecision: Serious Wide confidence intervals ;Publication bias: No serious

Length of stay in ahospital or

institution or both


Risk of bias: No serious Inadequate/lack of blinding of outcome assessors, resulting inpotential for detection bias ;Inconsistency: Serious due to LOS calculated in different ways(acute hospital vs totalhospital, two trials recorded median rather than mean and in two trials SD had to be inferredfrom p value or from results of similar trials). The magnitude of statistical heterogeneity washigh, with I^:86-88 %. ;Indirectness: No serious


17 of 157

References

[1] Stroke Unit Trialists' Collaboration Organised inpatient (stroke unit) care for stroke. Cochrane Database of SystematicReviews 2013;(9): Journal

Imprecision: No seriousPublication bias: No serious



Intervention: Organised stroke unit care

Comparator: General medical wards

OutcomeTimeframe



General medicalwards

Organised stroke unitcare

Certainty ineffect


Summary


follow-upMedian follow-up

of 12 months

4 Important

Odds Ratio 0.81(CI 95% 0.69 - 0.94)



up of 12 months

233per 1000

197per 1000


HighTwo different

models of care(comprehensive

stroke ward,mixed assessment

or rehabilitationward) tended to be

more effectivethan general

medical ward care.However, for the

comparison ofrehabilitation

stroke wards ormobile team care

(peripateticservice) versus

general medicalwards there were

no statisticallysignificant

differences.


(stroke unit) care weremore likely to survive

than those receiving carein general medical wards.


care by the end

Odds Ratio 0.78(CI 95% 0.68 - 0.89)


404per 1000

346per 1000

HighTwo different





18 of 157

http://dx.doi.org/10.1002/14651858.CD000197.pub3

of scheduledfollow-up

Median follow-upof 12 months

4 Important


up of 12 months











differences.

return home than thosereceiving care in general

medical wards.



Median follow-up12 months

7 Critical

Odds Ratio 0.79(CI 95% 0.68 - 0.9)



up of 12 months

615per 1000

558per 1000


HighTwo different











differences.



regain independencethan those receiving carein general medical wards.

Length of stay(days) in ahospital orinstitution

Median follow-up12 months

4 Important



Follow up Median follow-up of 12 months



LowInterpretation of

length of stay datawas complicated

by substantialheterogeneity.

There was no evidence ofa systematic increase in

length of stay.



19 of 157

References



up


Risk of bias: No serious Inadequate/lack of blinding of outcome assessors, resulting inpotential for detection bias ;Inconsistency: No seriousIndirectness: No seriousImprecision: No seriousPublication bias: No serious



up







Length of stay(days) in a hospital

or institution


Risk of bias: No serious Inadequate/lack of blinding of outcome assessors, resulting inpotential for detection bias ;Inconsistency: Serious The magnitude of statistical heterogeneity was high, with I^2:70 %. ;Indirectness: No seriousImprecision: No seriousPublication bias: No serious



Intervention: Different systems of organised care: acute stroke ward

Comparator: Alternative service

OutcomeTimeframe



Alternative service Different systems oforganised care: acute

stroke ward

Certainty ineffect


Summary


20 of 157



follow-up3 months - 1 year

4 Important

Odds Ratio 0.49(CI 95% 0.04 - 5.92)


(Randomized controlled)Follow up 3 months - 1

year

243per 1000

136per 1000


LowOverall, acute

(monitoring) unitsdid not havestatisticallysignificant

different odds ofdeath when

compared withacute (non-

intensive) units.

There is probably little orno difference in survival

between patients cared inacute stroke ward and

those cared in alternativeservice.



follow-up3 months - 1 year

4 Important

Odds Ratio 0.88(CI 95% 0.32 - 2.39)



year

429per 1000

398per 1000


LowOverall, acute


different odds ofdeath or requiringinstitutional carewhen comparedwith acute (non-intensive) units.

There is probably little orno difference in survival

or institutional carebetween patients cared in

acute stroke ward andthose cared in alternative

service.


the end ofscheduled follow

up3 months - 1 year

4 Important

Odds Ratio 0.76(CI 95% 0.24 - 2.41)



year

486per 1000

418per 1000


LowOverall, acute


different odds ofdeath or

dependency whencompared with

acute (non-intensive) units.

There is probably little orno difference in survivalor dependency betweenpatients cared in acutestroke ward and those

cared in alternativeservice.

Length of stay(days) in ahospital orinstitution

3 months - 1 year

4 Important

Based on data from: 265patients in 2 studies.


year



Very LowInterpretation of

length of stay wascomplicate day

substantialheterogeneity.

There was no evidence ofa systematic increase in

the length of stay.



up

Intervention reference:Systematic reviewBaseline/comparatorreference: Control arm of

Risk of bias: No serious Inadequate/lack of blinding of outcome assessors, resulting inpotential for detection bias ;Inconsistency: Serious The magnitude of statistical heterogeneity was high, with I^2:80%. ;Indirectness: No serious


21 of 157

References


reference used forintervention

Imprecision: Serious Wide confidence intervals, Low number of patients ;Publication bias: No serious



up


Risk of bias: No serious Inadequate/lack of blinding of outcome assessors, resulting inpotential for detection bias ;Inconsistency: Serious The magnitude of statistical heterogeneity was high, with I^2:64 %. ;Indirectness: No seriousImprecision: Serious Low number of patients, Wide confidence intervals, Wide confidenceintervals ;Publication bias: No serious


end of scheduledfollow up


Risk of bias: No serious Inadequate/lack of blinding of outcome assessors, resulting inpotential for detection bias ;Inconsistency: Serious The magnitude of statistical heterogeneity was high, with I^2:71 %. ;Indirectness: No seriousImprecision: Serious Wide confidence intervals, Low number of patients ;Publication bias: No serious


or institution


Risk of bias: No serious Inadequate/lack of blinding of outcome assessors, resulting inpotential for detection bias ;Inconsistency: Very Serious The magnitude of statistical heterogeneity was high, withI^2:92%. ;Indirectness: No seriousImprecision: Serious Wide confidence intervals, Low number of patients ;Publication bias: No serious



Intervention: Different systems of organised care: comprehensive stroke ward

Comparator: Alternative service (mobile stroke team)

OutcomeTimeframe



Alternative service(mobile stroke team)

Different systems oforganised care:comprehensive

stroke ward

Certainty ineffect


Summary


22 of 157



follow-up12 months

4 Important

Odds Ratio 0.35(CI 95% 0.19 - 0.65)


(Randomized controlled)Follow up 12 months

224per 1000

92per 1000


LowOrpington 2000 is

the only trial inthis analysis and

these resultsrequire furtherconfirmation.

Stroke unit care incomprehensive strokeward (providing acute

care and rehabilitation)may reduce death as

compared to admission togeneral wards where care

was provided by mobilestroke team.



follow-up12 months

4 Important

Odds Ratio 0.4(CI 95% 0.23 - 0.68)



296per 1000

144per 1000





Patients who receiveorganised stroke unit

care in comprehensivestroke ward (providing

acute care andrehabilitation) may be

more likely to survive andreturn home as compared

to admission to generalwards where care was

provided by mobilestroke team.


the end ofscheduledfollow-up12 months

4 Important

Odds Ratio 0.73(CI 95% 0.46 - 1.14)



480per 1000

403per 1000





Patients who receiveorganised stroke unit

care in comprehensivestroke ward (providing

acute care andrehabilitation) may be

more likely to survive andregain independence as

compared to admission togeneral wards where care

was provided by mobilestroke team.

Length of stay(days) in ahospital orinstitution12 months

4 Important



29.5days (Mean)

32days (Mean)

Difference: SMD 0.07 more( CI 95% 0.16 fewer - 0.3 more )




More research isrequired in this area.



up

Intervention reference:Systematic review [1] withincluded studies:Orpington 2000,Baseline/comparatorreference: Control arm of

Risk of bias: No seriousInconsistency: No seriousIndirectness: No seriousImprecision: No seriousPublication bias: No serious


23 of 157

References





up

Intervention reference:Systematic review [1] withincluded studies:Orpington 2000,Baseline/comparatorreference: Control arm ofreference used forintervention







or institution




Population: Organised inpatient (stroke unit) care for stroke

Intervention: Different systems of organised care: rehabilitation stroke ward

Comparator: Alternative service (mixed rehabilitation ward)

OutcomeTimeframe



Alternative service(mixed rehabilitation

ward)

Different systems oforganised care:

rehabilitation strokeward

Certainty ineffect


Summary


24 of 157



follow-upUp to 1 year

4 Important

Odds Ratio 0.51(CI 95% 0.29 - 0.9)



238per 1000

137per 1000


LowThe numbers were

small and nodefinite

conclusions couldbe drawn.

There was a pattern ofimproved outcomes in

the stroke rehabilitationward with statistically

significant fewer deaths.



follow-upUp to 1 year

4 Important

Odds Ratio 0.71(CI 95% 0.46 - 1.09)



506per 1000

421per 1000


LowThe numbers were




the stroke rehabilitationward with non-significanttrend for fewer patientswith the composite end

points of death orrequiring institutional

care



Up to 1 year

4 Important

Odds Ratio 0.8(CI 95% 0.45 - 1.42)



829per 1000

795per 1000


LowThe numbers were




the stroke rehabilitationward with a statisticallynon-significant trend forfewer patients with thecomposite end points of

death or dependency.

Length of stay(days) in ahospital orinstitution12 months

4 Important



(n/a) (n/a)

Difference: SMD 0.22 more( CI 95% 0.61 fewer - 1.05 more )

Very LowInterpretation of

length of stay datawas complicated

by substantialheterogeneity.

There was no evidence ofsystematic increase in

length of stay.



up

Intervention reference:Systematic review [1] withincluded studies: Dover1984 (MRW), Nottingham1996 (MRW), Orpington1993 (MRW),Baseline/comparatorreference: Control arm ofreference used forintervention

Risk of bias: No serious Inadequate/lack of blinding of outcome assessors, resulting inpotential for detection bias ;Inconsistency: No serious ;Indirectness: No seriousImprecision: No serious Low number of patients ;Publication bias: No serious


Intervention reference:Systematic review [1] with

Risk of bias: No serious Inadequate/lack of blinding of outcome assessors, resulting inpotential for detection bias ;


25 of 157

References



up

included studies: Dover1984 (MRW), Nottingham1996 (MRW), Orpington1993 (MRW),Baseline/comparatorreference: Control arm ofreference used forintervention

Inconsistency: No seriousIndirectness: No seriousImprecision: No serious Low number of patients ;Publication bias: No serious




Risk of bias: No serious Inadequate/lack of blinding of outcome assessors, resulting inpotential for detection bias ;Inconsistency: No seriousIndirectness: No seriousImprecision: No serious Low number of patients ;Publication bias: No serious


or institution


Risk of bias: No serious Inadequate/lack of blinding of outcome assessors, resulting inpotential for detection bias ;Inconsistency: Serious The magnitude of statistical heterogeneity was high, with I^2:92 %. ;Indirectness: No seriousImprecision: No seriousPublication bias: No serious


Population: Continuous versus intermittent physiological monitoring for acute stroke

Intervention: Continuous monitoring

Comparator: Intermittent monitoring of physiological variables

Summary

Ciccone et al (2013) [4] conducted a Cochrane review assessing whether continuous monitoring of physiological variablesaffected patients' prognosis of mortality or disability. Three studies were included (N = 354), including two randomised controlledtrials and one quasi-RCT where patients were allocated to continuous or intermittent monitoring based on the availability of beds.Continuous monitoring was associated with decreased death and disability at 3 months (OR 0.27, 95% CI 0.13 to 0.56), as well as


26 of 157


a non-significant reduction in all-cause mortality. However, the decrease in death and disability was non-significant whenexcluding the quasi-RCT with high risk of bias. Cardiac complications were also detected significantly more often, butcomparisons of other outcomes such as dependency, vascular death, and neurological complications showed no significantdifferences.

OutcomeTimeframe



Intermittentmonitoring ofphysiological

variables

Continuousmonitoring

Certainty ineffect


Summary


the end ofscheduled follow

upDischarge to 3

months

6 Important

Odds Ratio 0.27(CI 95% 0.13 - 0.56)


(Randomized controlled)Follow up Discharge to 3

months

469per 1000

193per 1000


LowThe evidence waslow because the

trial whichcontributed most

to the primaryoutcome (Cavallini

2003) was nottruly randomised

as participantswere allocated to a

conventionalstroke unit or to a

stroke unit withcontinuous

monitoring purelyon the basis of bedavailability, therewas no long-termfollow up and it is

not certain thatthe assessment of

outcomes wasblinded. If this

study is removedfrom the meta-

analysis the resultis no longerstatistically

significant (OR0.32, 95% CI 0.06

to 1.63), withconsistent

heterogeneitybetween the tworemaining studies(I2 = 67%, 95% CI

93% to 44%).

Continuous monitoringsignificantly reduced

death and disability atthree months or at

discharge but theseresults depended on onestudy at high risk of bias.

Cardiaccomplications

Discharge to threemonths

Odds Ratio 8.65(CI 95% 2.52 - 29.66)



17per 1000

130per 1000

Difference: 113 more per 1000

LowThe quality of theevidence was low.There was slightheterogeneity of

Continuous monitoringwas associated with a

significant increase in thedetection of cardiac

complications


27 of 157

6 Important

Follow up Discharge tothree months

( CI 95% 25 more - 322 more )

the studies, a smallnumber of studies,

small samplessizes and a high

risk of bias for thetrial that

contributed mostin terms of the

number ofparticipants

enrolled.

(arrhythmias, heartfailure, myocardial

infarction)

FeverDischarge to three

months

5 Important

Odds Ratio 2.17(CI 95% 1.27 - 3.7)


(Randomized controlled)Follow up Discharge to

three months

158per 1000

289per 1000

Difference: 131 more per 1000( CI 95% 34 more - 252 more )

LowThe quality of theevidence was low.There was slightheterogeneity of

the studies, a smallnumber of studies,

small samplessizes and a high

risk of bias for thetrial that

contributed mostin terms of the

number ofparticipants

enrolled.


significant increase in thedetection of fever.

Length of stay(days)

Discharge to threemonths

6 Important



three months

(n/a) (n/a)

Difference: MD 5.24 fewer( CI 95% 10.51 fewer - 0.03 more )

Very LowThere are a small

number of studiesand small samples

sizes. There wassubstantial

heterogeneityacross trials for

this outcome (I^2= 83%, 95% CI

94% to 49%; P =0.003) and if theVERITAS 2007

study wasremoved from the

analysis thereduction in

hospital stay withcontinuous

monitoring wasstatistically

significant (meandifference (MD)

-8.15 days, 95% CI-9.85 to -6.44)

without significantinconsistency (I2 =

0%, P = 0.54).


non-significant reductionin the number of days of

hospital stay.


28 of 157



end of scheduledfollow up

Intervention reference:Systematic review [4] withincluded studies: Cavallini2003, Sulter 2003,VERITAS 2007,Baseline/comparatorreference: Control arm ofreference used forintervention

Risk of bias: Serious Inadequate sequence generation/ generation of comparable groups,resulting in potential for selection bias, Inadequate concealment of allocation duringrandomization process, resulting in potential for selection bias ;Inconsistency: No seriousIndirectness: No seriousImprecision: No seriousPublication bias: No serious

Cardiaccomplications


Risk of bias: Very Serious Inadequate sequence generation/ generation of comparablegroups, resulting in potential for selection bias, Inadequate concealment of allocation duringrandomization process, resulting in potential for selection bias ;Inconsistency: No seriousIndirectness: No seriousImprecision: Serious Low number of patients ;Publication bias: No serious

Fever


Risk of bias: Very Serious Inadequate sequence generation/ generation of comparablegroups, resulting in potential for selection bias, Inadequate concealment of allocation duringrandomization process, resulting in potential for selection bias ;Inconsistency: No seriousIndirectness: No seriousImprecision: Serious Low number of patients ;Publication bias: No serious

Length of stay(days)


Risk of bias: Very Serious Inadequate sequence generation/ generation of comparablegroups, resulting in potential for selection bias, Inadequate concealment of allocation duringrandomization process, resulting in potential for selection bias ;Inconsistency: Serious The magnitude of statistical heterogeneity was high, with I^2:83 %. ;Indirectness: No seriousImprecision: Serious Low number of patients ;Publication bias: No serious


29 of 157

Practice Statement

1. All people with stroke should be admitted directly to a stroke unit (preferably within three hours of stroke onset).

2. If people with suspected stroke present to non-stroke unit hospitals, transfer protocols should be developed and used to guide urgent

transfers to the nearest stroke unit hospital.

3. Where transfer is not feasible, sSmaller isolated hospitals should managerconsider stroke services in a manner that adheres as closely

as possible to the criteria for stroke unit care. Where possible, patients should receive care in geographically discrete units.If people with

suspected stroke present to non-stroke unit hospitals, transfer protocols should be developed and used to guide urgent transfers to the

nearest stroke unit hospital.

Practical Info

All patients should be informed about options for transfer and the benefits of transport to a specialist stroke service.


All acute stroke services should use standarised protocols to manage physiological status and prevent complications. (Middleton et al

2011 [97]).

Practical Info

In Quality in Acute Stroke Care (QASC) study, monitoring and prompt treatment of hyperglycaemia, fever and swallowing dysfunctionwere critical in improving healthcare process and patient outcomes. For details on the management of these complications, refer to thesection Glycaemic therapy, Pyrexia management, and Dysphagia.

Key Info

Rationale

Quality in Acute Stroke Care (QASC) study provided evidence that an acute protocol aiming to ensure monitoring and prompt treatmentof common complications fever, hyperglycaemia and swallowing reduced death and dependency in patients in stroke units. Furthermore,it is likely that patients would want to receive this best standard care. Therefore it should be provided to all stroke patients.

Benefits and harms

A multidisciplinary, nurse-initiated treatment protocol for the management of fever, hyperglycaemia, and swallowing dysfunctiondemonstrated significant reduction of death and dependency at 90 days (157 less per 1000, number needed to treat 6.4d), andimprovement of physical health (3.4 higher on SF-36 physical health score). Functional independence measured with Barthel Indexalso indicated a non-significant trend of improvement.


Quality of evidence

The quality of evidence is moderate as only one study albeit a large multi-centre randomised controlled trial with high methodologicalquality.

Moderate


It is expected that patients would want to receive this protocol shown to improve their outcomes.



30 of 157



Intervention: Acute nursing intervention

Comparator: Control

Summary

The Quality in Acute Stroke Care (QASC) study conducted by Middleton et al (2011) [97] was a single-blind cluster randomisedtrial, assessing the benefits of evidence-based treatment protocols in acute stroke units. The Fever, Sugar, Swallowing (FeSS)intervention involved temperature monitoring, monitoring of blood glucose and dysphagia assessment and was aimed atpromoting prompt nursing assessment and bedside treatment. The results showed a significant reduction in death or dependencyat 90 days (modified Rankin Scale scores >= 2), with an adjusted absolute risk reduction of 15.7%. The intervention group alsoshowed higher rates of functional independence, both when independence was classified as a Barthel Index score >= 60 or >= 95,although the difference was non-significant. Other outcomes suggested improved processes of care in the intervention strokeunits, with significantly reduced temperatures and blood glucose, and higher proportions of swallowing screening. Patients withsevere strokes may have been underrepresented due to the exclusion of patients receiving palliation only, but in other respectsthe study was high quality and provides a high degree of certainty about the observed results.

OutcomeTimeframe



Control Acute nursingintervention

Certainty ineffect


Summary

Death ordependency90 days after

admission

9 Critical

n/a


(Randomized controlled)Follow up 90 days

577per 1000

420per 1000


ModerateDue to serious

imprecision

The FeSS protocol foracute stroke care

probably decreases deathor dependency

Functionalindependence(Barthel Index

>= 95)

8 Critical

n/a



600per 1000

695per 1000

Difference: 95 more per 1000( CI 95% 5 fewer - 195 more )


imprecision


probably has little or nodifference on functionalindependence (Barthel

Index >= 95)

FunctionalIndependence(Barthel Index

>= 60)

8 Critical

n/a



898per 1000

923per 1000



imprecision

The FeSS protocol foracute stroke care has

little or no difference onfunctional independence

(Barthel Index >= 60)


31 of 157

Physical health90 days after

admission

7 Critical

Measured by: SF-36Physical health score

High betterBased on data from:1,009 patients in 1


Follow up 90 days

points (n/a) points (n/a)

Difference: MD 3.4 more( CI 95% 1.2 more - 5.5 more )


imprecision


improves physical health

Mental health90 days after

admission

7 Critical

Measured by: SF-36 Mentalhealth scoreHigh better



Follow up 90 days

points (n/a) points (n/a)

Difference: MD 0.5 more( CI 95% 1.9 fewer - 2.8 more )


imprecision

The FeSS protocol foracute stroke care has

little or no difference onmental health


Death ordependency

Intervention reference:Primary study [97]middleton_2011_implement,Baseline/comparatorreference: Control arm ofreference used forintervention

Risk of bias: No seriousInconsistency: No seriousIndirectness: No serious Excluded palliative patients so may have under-representedsevere stroke patients. ;Imprecision: Serious Only data from one study ;Publication bias: No serious

Functionalindependence

(Barthel Index >=95)



FunctionalIndependence



Risk of bias: No seriousInconsistency: No seriousIndirectness: No seriousImprecision: Serious Only data from one study ;Publication bias: No serious

Physical health




32 of 157

References

[97] Middleton S, McElduff P, Ward J, Grimshaw JM, Dale S, D'Este C, Drury P, Griffiths R, Cheung NW, Quinn C, Evans M,Cadilhac D, Levi C Implementation of evidence-based treatment protocols to manage fever, hyperglycaemia, and swallowingdysfunction in acute stroke (QASC): a cluster randomised controlled trial.. Lancet (London, England) 2011;378(9804):1699-706-Pubmed Journal

Mental health

Intervention reference:Systematic reviewmiddleton_2011_implement,Baseline/comparatorreference: Control arm ofreference used forintervention



33 of 157

http://www.ncbi.nlm.nih.gov/pubmed/21996470

http://dx.doi.org/10.1016/S0140-6736(11)61485-2

3 - Palliative care

Eleven per cent of acute stroke patients admitted to hospital die in hospital (?% within seven days) (Stroke Foundation 2014) andapproximately 20% die as a result of the stroke in the first 30 days (Thrift et al 2000 [19]).

Practical end-of-life issues, such as the use of medical power of attorney and advance care directives, should be discussed. Organ donationmay be sensitively raised if appropriate. Issues of bereavement may become part of the responsibility of the stroke team.


Stroke patients and their families/carers should have access to specialist palliative care teams as needed and receive care consistent with

the principles and philosophies of palliative care.

Key Info

Benefits and harms

Gade et al. (2008 ) [9] (N=512) showed that interdisciplinary palliative care could improve advanced directives ( n u m b e r n e e d e d to t r e a t t o b e n e f i t 7 . 5 ) ; decrease ICU admissions; improve patient satisfaction and communication with providers and increaselength of hospice stay. No harms were reported.


Quality of evidence

The evidence is considered moderate due to confidence intervals not being reported which made the range of possible benefit hard todetermine.

Moderate


The qualitative studies by Payne et al . (2010 ) [14]; Burton C, and Payne S (2012) [13] and; de Boer et a l . l a (2015) [15] discuss theneed for palliative care services not to focus exclusively on end of life care but also to support quality of life for patients who have hada stroke and are likely to have a poor outcome and/or die in the acute phase of care.

The studies suggested that the significant advances made to implement evidence of rapid neurological assessment, specialistmanagement and organised stroke services will mean that there will an increasing need for patients to have access to specialistpalliative care services when needed and for all staff to be appropriately trained in palliative/supportive care.

Although these studies were undertaken mainly in the UK they would have direct applicability to the stroke unit care model inAustralia and the needs of patients and their families/carers in relation to palliative care.

Whilst the evidence of patient’s views on palliative care is understandably limited it is clear that from a patient’s perspective themanagement of physical symptoms and psychological distress when the outcome of their stroke is likely to lead to major disability/death is i s appropriate and needed.

Blacquiere et al. (2013) [16] , a Canadian study quantified the satisfaction with palliative care of families of patients who had died fromstroke. Overall their satisfaction was high (9.04 out of 10) with most satisfaction about decision making but least about emotionalneeds being met. There was less satisfaction about the control of individual symptoms and provision of adequate information. Themost contentious area was the cessation of artificial hydration and feeding.

Although none of the studies directly assess whether families wanted palliative care for the patient following a stroke there is supportfor the provision of this care when needed. These limited studies identified an expressed desire from families for the patient to be painfree and not suffering emotional distress. It is also clear that the satisfaction ratings support the view that the families valued thepalliative care they received although they thought it could be improved.

It is not likely that the values and preferences in the Australian context would differ significantly.



34 of 157

Rationale

Gade et al. (2008) [9] demonstrated that multidisciplinary palliative care teams reduced hospital admissions and increased decisionmaking (number of advanced care directives). They also improved communication and patient satisfaction slightly. A number of studies(both qualitative and quantitative) reported that the management of physical symptoms and psychological distress when the outcome ofstroke is likely to lead to major disability/death is is appropriate and needed. It was also reported that there is a need to not focusexclusively on end of life care but also to support quality of life for patients who have had a stroke and are likely to have a poor outcomeand/or die in the acute phase of care.


Gade et al. (2008) [9] reported significantly lower total health costs for patients randomized to inpatient palliative care services.Mean total costs were USD$14,486 in the IPCS group and USD$21,252 in the usual care group, with the difference driven by lowerhospital readmission costs ($6,421 per patient for IPCS and $13,275 for usual care). IPCS patients also had significantly fewer ICUstays on readmission.




Intervention: Interdisciplinary palliative care

Comparator: Usual care

Summary

Gade et al (2008) [9] carried out a multicenter randomized controlled trial (N = 517) to assess the impact of an interdisciplinarypalliative care service (IPCS) compared to usual hospital care. The IPCS care teams included palliative care nurses and physicians,social workers and chaplains. Primary outcomes were "symptom control, levels of emotional and spiritual support,patient satisfaction, and total health services costs at 6 months postindex hospitalization". Patients receiving IPCS care reportedhigher satisfaction on the Care Experience scale and Doctors, Nurses/Other Care Providers Communication scale . Total costswere also lower in the IPCS group, with a mean 6-month saving of $4,855 (USD) per patient. IPCS patients were also more likely tohave completed advanced directives by the time of hospital discharge, and had lower numbers of ICU admissions. There were nodifferences in survival or symptom control between the groups.

Creutzfield et al (2012) [12] conducted a narrative literature review investigating the palliative care needs of stroke survivors. Thereview included evidence for central poststroke pain, hemiplegic shoulder pain, painful spasticity, fatigue, incontinence, post-stroke seizures, sexual dysfunction, sleep-disordered breathing, depression and emotionalism. The authors also reviewed the roleof caregivers and ways to support them. The literature search was conducted using PubMed; searching from 1995 and limited toclinical practice guidelines and RCTs. Outcomes of interest included:

Pain: specifically central post-stroke pain (CPSP) and hemiplegic shoulder pain (HSP). One study (N=15) found the tricyclicantidepressant (TCA) amitriptyline to be effective in CPSP with other TCA's and selective noradrenergic receptor inhibitors(SNRIs) showing effectiveness for neuropathic pain. The anticonvulsant Lamotrigine was moderately effective in 30 patients withCPSP. For HSP the authors found that a shoulder sling during ambulation may support the arm to reduce pain and prevent upperextremity trauma. Promising interventions requiring further study include IM Botox-A, intra-articular steroid injections andneuromuscular electrical stimulation.

Psychological outcomes reviewed included post-stroke depression (PSD), anxiety and emotionalism. The efficacy of medicationsto prevent PSD is unclear however pharmacologic treatment of PSD was found to lead to a reduction in various measures ofdepression but it was unclear what effect they have on functional outcomes. Adverse events were common and included centralnervous system events (confusion, sedation, tremor) and GI effects (constipation, diarrhoea). Controlled trials on PSD were


35 of 157

limited to TCAs and SSRIs and they found that while TCAs are effective in reducing depression, their cholinergic side effects limittheir clinical usefulness, especially in older, frail patients with vascular disease. The data for SSRIs was mixed however the safetyprofile was more favourable making them the drug of choice. One study suggested that the SSRI citalopram may be more effectivein “anxious depressed” (agitated, irritable) patients, whereas the noradrenergic drug reboxetine may be more effective in“retarded depressed” (mentally and physically slowed down) patients. While psychological, “talking” interventions (mostlybehavioral interventions: identifying symptoms and causes of depression, and identifying and planning pleasant activities)seem promising, their benefit is not yet convincing,and their use should be tailored individually. Anxiety may accompanydepression so for this reason antidepressant medications (e.g., citalopram) may be effective for generalised anxiety or panicpattern symptoms in this setting. If anxiety is severe and if the lifespan is limited, however, benzodiazepines are the drugs ofchoice. No drugs were recommended for emotionalism at this time.

Social outcomes reviewed included:Care giving and receiving. Women, younger caregivers, those with poor physical health and those caring for patients with severecognitive, behavioural and emotional changes are at highest risk of caregiver burn-out. Support programs should focuson increasing self-efficacy, active coping strategies and social support. If necessary, referrals should be made to appropriateservices that meet identified social needs and promote access to care, transportation, rehabilitation, medications, counselling,community resources and equipment. Common fears specific to stroke caregivers are caused by the uncertainty of prognosis withthe fear of another stroke, and the feeling of abandonment, especially when their loved one is unable to communicate. Caregiver’sneeds include information provision, managing emotions, social support, health maintenance and practical problem solving.Training caregivers in their new role has been shown to reduce perceived and actual burden while improving psychosocialoutcomes in both caregivers and patients. Over 90% of caregivers also reported that their experience as a stroke caregiver hadincreased their appreciation of life. The authors recommend consultation with a local social worker familiar with resources in thepatient’s community to ensure that all opportunities are explored.

OutcomeTimeframe



Usual care Interdisciplinarypalliative care

Certainty ineffect


Summary

Advancedirectives

At discharge

n/a


(Randomized controlled)Follow up At discharge

778per 1000

911per 1000

Difference: 133 more per 1000

ModerateThe difference was

significant (chi-squared test) but

confidenceintervals were notreported. Due to

seriousimprecision

Interdisciplinarypalliative care probablyincreases completion of

advance directives

ICU admissionsn/a



96per 1000

52per 1000


significantaccording to a chi-squared test, but

confidenceintervals were notreported. Due to

seriousimprecision

Interdisciplinarypalliative care probably

decreases ICUadmissions


36 of 157

Patientsatisfaction -

careenvironment2 weeks after

discharge

Measured by: MCOHPQPlace of Care

environment scaleScale: 0-10 High better


(Randomized controlled)Follow up Within 2 weeks

of discharge

6.4(Mean)

6.8(Mean)

Difference: MD 0.4 moreCI 95%


significant but noconfidence

intervals werereported. Due to

seriousimprecision


improves patientsatisfaction with the care

environment slightly

Patientsatisfaction -

communicationwith providers

2 weeks afterdischarge

Measured by: Doctors,Nurses / Other Care

Providers -Communication scale

Scale: 0-10 High betterBased on data from: 341

patients in 1 studies.(Randomized controlled)

Follow up Within 2 weeksof discharge

7.4(Mean)

8(Mean)

Difference: MD 0.6 moreCI 95%




seriousimprecision


improves patientsatisfaction regardingcommunication with

providers slightly

Hospice lengthof stay

Measured by: Days inhospice



12days (Median)

24days (Median)

Difference: 12 moreCI 95%




seriousimprecision


increases hospice lengthof stay


Advance directives

Intervention reference:Primary studyBaseline/comparatorreference: Control arm ofreference used forintervention

Risk of bias: No seriousInconsistency: No seriousIndirectness: No seriousImprecision: Serious Only data from one study, confidence intervals not reported so range ofpossible benefit hard to determine ;Publication bias: No serious

ICU admissions



Patient satisfaction- care environment




37 of 157

Practice Statement

1. An accurate assessment of prognosis or imminent death should be made for patients with severe stroke or those who are

deteriorating.

2. A pathway for stroke palliative care can be used to support stroke patients and their families/carers and improve care for people dying

after stroke.

Key Info

References

[9] Gade G, Venohr I, Conner D, McGrady K, Beane J, Richardson RH, Williams MP, Liberson M, Blum M, Della Penna R Impact ofan inpatient palliative care team: a randomized control trial.. Journal of palliative medicine 2008;11(2):180-90- Pubmed Journal

[12] Creutzfeldt CJ, Holloway RG, Walker M. Symptomatic and palliative care for stroke survivors. Journal of General InternalMedicine 2012;27(7):853-860 Journal

Patient satisfaction- communication

with providers



Hospice length ofstay




The qualitative studies by Payne et al (2010); Burton C, and Payne S (2012) and; de Boer et la (2015) discuss the need for palliativecare services not to focus exclusively on end of life care but also to support quality of life for patients who have had a stroke and arelikely to have a poor outcome and/or die in the acute phase of care.

The studies suggested that the significant advances made to implement evidence of rapid neurological assessment, specialistmanagement and organised stroke services will mean that there will an increasing need for patients to have access to specialistpalliative care services when needed and for all staff to be appropriately trained in palliative/supportive care. The Burton and Paynestudy, that included a systematic review of international literature on palliative care within stroke (7 studies, 4 of which were from theUK), identifies the complexity and challenges of integrating palliative care into the acute stroke pathway. This study that aimed toinvestigate the interface of stroke and palliative care from the perspective of the patients, family members and stroke service staff isnot conclusive but suggests that further investigation of the practical difficulties such as the stroke unit environment, prognosticaccuracy of the information communicated, timing and shared decision making, as usually the patient is unable to participate and thenext of kin is required to make the decision about treatment. The de Boer et al qualitative study focuses specifically on this processand found relatives’ experienced difficulties related to having to make choices under pressure, not feeling able to legitimately make adecision for someone else and coping with unexpected changes.



38 of 157


http://dx.doi.org/10.1089/jpm.2007.0055

http://dx.doi.org/http://dx.doi.org/10.1007/s11606-011-1966-4

Rationale

Mortality after stroke is not insignificant. A previous systematic review (7 trials) showed that carers of stroke patients have differentneeds to those involved in specialist palliative care in cancer. They require more support, particularly as they are likely to be older and inpoor health, and caring for their family members in difficult circumstances, often unsupported (Stevens et al., 2007, [17]).

An observational study was identified that developed and implemented a care pathway for palliative care in acute stroke. The studyreported improved processes of care based on national standards compared to care provided prior to the pathway (Jack et al., 2004, [18]).

Although these studies were undertaken mainly in the UK they would have direct applicability to the stroke unit care model inAustralia and the needs of patients and their families/carers in relation to palliative care. The change in stroke services from ahistorically nihilistic approach and passive care to services that reflect neurological urgency and optimism are likely to make it moredifficult for clinicians and the patient’s family/carers to cope with a poor outcome. Therefore the need for palliative/supportive carewithin the acute stroke pathway when needed is critical for the patient and their families /carers. Although the way in which this isprovided is not clear, all the studies support this view.

Payne et al (2010) included qualitative information from a small sample (28) of patients with an acute stroke who had high and lowdisability. Although the interviews did not focus on palliation the patients indicated a wish to be involved in decisions in about theirmedical care, in particular when facing uncertainty about their recovery following the stroke. They also expressed views suggestingthat they wanted to know that they were receiving the correct treatment but had a fatalist perspective on their likely outcome andwere perhaps more concerned about partners and families. The study cited previous studies that found that the physical symptomburden and psychological distress for patients with a poor prognosis was comparable to cancer patients and they conclude from theirdata that patients who experience acute stroke have needs and concerns that require compassionate end of life care before thepatient can be described as dying.Whilst the evidence of patient’s views on palliative care is understandably limited it is clear that from a patient’s perspective themanagement of physical symptoms and psychological distress when the outcome of their stroke is likely to lead to major disability/death is is appropriate and needed.

Payne et al (2010) and the subsequent study in 2012 by Burton and Payne use qualitative data from semi structured interviews thatfocused on communication and facing uncertainty and end of life preferences. The studies have limitations due to the small sample(25 family members) but provide some support from families for palliative care to be provided in the acute stroke unit. In generalfamilies expressed concern about the stroke patient suffering any form of distress, they wanted a “peaceful, pain free and dignifiedending”, they wanted good communication about the likely outcome and to be involved in decisions about active or passiveinterventions. The views of the families suggest that they wanted compassionate palliative care to be provided.

Blacquiere et al (2013), a Canadian study quantified the satisfaction with palliative care of families of patients who had died fromstroke. They measured a range of parameters covering physical and psychological care for 15 families of patients who had an ischemicstroke, intracerebral or subarachnoid hemorrhage. Overall their satisfaction was high (9.04 out of 10) with most satisfaction aboutdecision making but least about emotional needs being met. There was less satisfaction about the control of individual symptoms andprovision of adequate information. The most contentious area was the cessation of artificial hydration and feeding.

Although none of the studies directly assess whether families wanted palliative care for the patient following a stroke there is supportfor the provision of this care when needed. These limited studies identified an expressed desire from families for the patient to be painfree and not suffering emotional distress. It is also clear that the satisfaction ratings support the view that the families valued thepalliative care they received although they thought it could be improved.

It is not likely that the values and preferences in the Australian context would differ significantly.


39 of 157

4 - Reperfusion therapy

4.1 - Thrombolysis

From 2010 guidelines:

Access to thrombolysis remains low in Australia (~3% of all ischaemic stroke patients).11 However some Australian centres have achieved

thrombolysis rates of 20%.26 Only 39% of patients arrive within 4.5 hours and pre-hospital delays (particularly time to seek medical help)

remain one of the main challenges.11 Intravenous rt-PA was licensed by the Australian Therapeutic Goods Administration for use within threehours in acute ischaemic stroke in October 2003. Intra-arterial and mechanical clot retrieval are discussed separately (see 4.2Neurointervention).

One updated Cochrane review (26 RCTs) examined four different thrombolytic agents: rt-PA, streptokinase, recombinant pro-urokinase, and

urokinase, 56% of the data coming from trials of rt-PA.12 Only 0.5% of the data came from patients over 80 years of age. Thrombolysis in alltrials and all agents combined resulted in a significant reduction in the combined end-point of death or disability (OR 0.81, 95% CI 0.73–0.90).Thrombolysis (all agents pooled) showed a net benefit, but is associated with a risk of intracerebral haemorrhage (ICH) at the end of three orsix month follow-up (OR 3.49, 95% CI 2.81–4.33). The effect of rt-PA on death or dependency was similar whether given within three hours(OR 0.69, 95% CI 0.44–1.09) or later than three hours after stroke (OR 0.88, 95% CI 0.73–1.06), although there is a strong trend towardsbetter outcome with earlier intervention (I2=25%, p=0.09). There were no differences between agents in terms of symptomatic intracranialhaemorrhage (sICH), death or dependency but this conclusion is based on indirect comparisons and heterogeneity was noted. More robustdata are needed before agents other than rt-PA can be recommended. Concurrent antithrombotic therapy increased adverse events with theodds of death by the end of follow-up found to increase (antithrombotic therapy within 24 hours of thrombolysis OR 1.92, 95% CI 1.43-2.57;no antithrombotic drugs within the first 10–14 days OR 0.89, 95% CI 0.58-1.37). These conclusions are however based on non-randomisedcomparisons. The review concluded that thrombolytic therapy appears most beneficial if provided in experienced centres in highly selected

patients. Widespread use in routine clinical practice in non-organised stroke care is not recommended.12

Another pooled analysis (six rt-PA trials) confirmed that intervention with rt-PA had a clear net benefit in reducing the odds of death or

dependency if given within three hours.220 Odds of functional independence were 30% greater with a 12% absolute difference between the

rt-PA intervention group and controls.220 Using data from the same six RCTs included in the pooled analysis, the NNT/NNH estimates are 3.6/65 (0–90 minutes), 4.3/38 (91–180 minutes), 5.9/30 (181–270 minutes), and 19.3/14 (271–360 minutes).221 The ECASS III RCT (included inthe updated Cochrane review 12) found rt-PA to be effective when provided up to 4.5 hours after stroke onset (OR 1.34, 95% CI 1.02–1.76).There was a significant increase in sICH (2.7% vs 0.3%) but no significant effect on deaths (6.7% vs 8.2%).222 A systematic review (seven trialsincluding the ECASS III study) confirmed that rt-PA given 3–4.5 hours after stroke onset is associated with an increased chance of favourable

outcome (OR 1.31, 95% CI 1.10–1.56) with no significant difference in mortality (OR 1.04, 95% CI 0.75–1.43) compared to placebo.223

Phase IV studies and large registries have shown rt-PA to be as safe (with often lower adverse events reported) and effective in clinical

practice as in the major trials.224-226 Careful patient selection, strict protocol adherence including close monitoring of patient vital signs

(particularly high blood pressure which is clearly associated with poor outcomes) 227, 228 audit and quality improvement processes are

strongly recommended for all centres delivering rt-PA.229

The available evidence shows that intravenous rt-PA therapy is beneficial for selected patients but should be delivered in well-equipped andskilled EDs and/or stroke care units with adequate expertise and infrastructure for monitoring, rapid assessment and investigation of acute

stroke patients.229 Collaboration between clinicians in pre-hospital emergency services, emergency medicine, neurology and neuroradiologyis recommended to enable prompt identification of potentially eligible patients, expert patient selection and audit and quality improvement

initiatives.229 Models for improving access to rt-PA for rural and regional centres including telestroke and or transfer protocols urgently needto be developed and tested to ensure greater equity of services across Australia.

Advanced MR and CT imaging may help identify ischaemic but potentially viable brain tissue in patients considered for thrombolysis,particularly those presenting beyond the currently accepted maximum time window for rt-PA (4.5 hours). While thrombolysis based on MRI

selection has been shown to attenuate infarct growth 230, overall, advanced imaging has not been shown in RCTs to have any effect on patientoutcomes (see 3.3 Imaging).12, 231 Of the many observational studies, the largest registry (n=1210) reported that MRI selection significantlyreduced symptomatic intracranial hemorrhage (OR 0.52, 95% CI 0.27–1.0) compared to standard CT selection. Beyond three hours, MRIsignificantly predicted a favorable outcome (OR 1.47, 95% CI 1.02–2.12). Under three hours and for all secondary end-points, there was a

trend in favor of MRI-based selection over standard CT-based intervention.232


40 of 157

One systematic review (six RCTs and three non-randomised trials) found sonothrombolyis (ultrasound-assisted thrombolysis) increased thelikelihood of complete recanalisation for high-frequency applications compared to routine rt-PA alone (OR 2.99, 95% CI 1.70–5.25) and didnot increase the risk of sICH (OR 1.26, 95% CI 0.44–3.60). Low-frequency ultrasound led to higher rates of sICH compared to rt-PA alone

(35.7% vs 17.2%).233

The failure to implement stroke thrombolysis is an international problem but numerous studies have demonstrated rates of up to 20% areachievable. In Australia, new models of care need to be developed and assessed and it is likely that, given the international experience, ‘one-size-fits-all’ solutions will fail. Local and network interventions will need to be developed and evaluated. Such interventions may need toinclude telemedicine resources and training for regional and rural centres, systems-level coordination and changes, and appropriate numbersof trained acute stroke personnel with obvious implications for ongoing training and support. Given the potential risks of thrombolysis, thereis the potential for adverse outcomes with inappropriate use, and routine audit and ongoing quality improvement will be important to identifyproblem areas and local solutions.


• Intravenous alteplase (dose of 0.9 mg/kg , maximum o f 9 0 mg) in acute ischaemic stroke should be administered in patients with

potentially disabling ischaemic stroke meeting specific eligibility criteria (Wardlaw 2014 [20]; Emberson et al. 2014 [21]).

• Thrombolysis should commence as early as possible (within the first few hours) after symptom onset but may be used up to 4.5

hours after onset (Anderson et al. 2016 [23]).

Practical Info

Intravenous thrombolysis eligibility should be determined by an assessment of the balance of risk versus benefit in the individual patient."Potentially disabling ischemic stroke" does not require any particular threshold score to be achieved on the National Institutes of HealthStroke Scale. The NINDS tPA trial included patients with "measurable neurological deficit". For example, an isolated hemianopia (NIHSS 2)would qualify as potentially disabling and isolated dysphasia or significant hand weakness with minimal arm drift may also warranttreatment. Although some guidelines have recommended against treatment of "severe" stroke, individual patient data meta-analysisdemonstrated consistent benefit of treatment across the spectrum of stroke severity. [21] Rapidly improving clinical severity hassometimes been regarded as an exclusion from thrombolysis. However, these patients have a substantial risk of subsequent deteriorationand treatment should be considered if there is still a potentially disabling deficit or if imaging indicates a persisting vessel occlusion(Coutts et al., 2012, [33]).

Benefit from thrombolysis is strongly time dependent and so treatment should be commenced as early as possible within the 4.5 hourperiod after symptom onset. Commencing treatment beyond 4.5 hours has not been shown to be of net benefit. Ongoing trials usingimaging selection are evaluating thrombolysis in this group of patients.

Contraindications to thrombolysis generally relate to either systemic or intracerebral bleeding risk or potential alternative diagnosis of astroke mimic. Many proposed criteria were adopted directly from trial exclusion criteria and studies of "off label" thrombolysis havesuggested that some of these may not be well justified. Limitations on age have been imposed in some trials but meta-analysis of allalteplase trials has clearly demonstrated a treatment benefit of at least as great a magnitude in patients aged >80 compared to youngerpatients. The trial data did not demonstrate an increase in symptomatic intracerebral hemorrhage in the elderly. The patient's premorbidlevel of function rather than chronological age should be considered when deciding whether to treat.

Scenarios that have potentially mimicked stroke and therefore been regarded as exclusions from some trials include seizure at onset,hypoglycaemia and hyperglycaemia. Brain imaging can potentially overcome these diagnostic pitfalls by proving a diagnosis of ischaemicstroke. Seizure should not prevent thrombolysis if there is a vessel occlusion or perfusion lesion diagnostic of stroke and there has been nosignificant trauma as a result of the seizure. Hypoglycaemia should be corrected and, if symptoms remain and there is imaging evidence ofstroke, the patient can receive thrombolysis. Hyperglycaemia is a negative prognostic factor but, in the presence of confirmed diagnosis ofstroke, should not prevent thrombolysis and the hyperglycaemia should be treated in parallel.

Contraindications:• Acute intracranial hemorrhage• Extensive frank hypodensity on CT scan (greater than 1/3 of middle cerebral artery territory or equivalent)This should prompt reassessment of the stroke onset time. Subtle ischemic changes (loss of grey white differentiation) are not acontraindication but reflect irreversible injury.


41 of 157

• Active non-compressible systemic bleeding• Systemic coagulopathy (NB thrombolysis should not be delayed by coagulation testing unless there is clinical suspicion of coagulopathy)- platelet count <100,000mm3 (based on expert consensus)- INR >1.7, including warfarin use (based on limited observational data)- unfractionated heparin within 48 hours with an elevated APTT- low molecular weight heparin within 24 hours (including prophylactic doses)- direct oral anticoagulant use (e.g. apixaban, dabigatran, rivaroxaban) within 48 hours with abnormal coagulation parameters asappropriate to the particular medication, unless a specific reversal agent is available (see below)• Infective endocarditis (increased risk of symptomatic intracerebral hemorrhage)

Relative contraindications (careful consideration of risk and benefit required):• Severe uncontrolled high blood pressure: the standard recommendation based on expert consensus is to lower elevated bloodpressure to <185/110mmHg prior to thrombolysis and maintain this level. If blood pressure cannot be lowered then thrombolysisshould not be commenced.• Previous intracerebral haemorrhage (not including cerebral microbleeds on MRI)• Cranial or spinal surgery or major head trauma within 3 months (expert consensus)• Other major surgery or trauma within 14 days (expert consensus) - consider discussion with the surgeon involved• Recent gastrointestinal or genitourinary tract bleeding within 21 days (expert consensus)• Central nervous system intra-axial neoplasm (i.e. not meningioma)• Ischemic stroke within 3 months (consider the size of the previous infarct and severity of current stroke)

Other notes:Cervical artery (extra-cranial) dissection - available data suggest alteplase is safe in these patients

Thoracic aortic dissection - alteplase is contraindicated due to a high risk of death

Pregnancy - there is no known fetal toxicity related to alteplase but experience is limited. Uterine bleeding and fetal death is a potentialrisk. This needs to be balanced against the risk of the stroke and potential alternative endovascular treatment.

Menstrual bleeding is not regarded as a contraindication to thrombolysis but should be monitored in the first 24 hours.

Unruptured aneurysms have not been demonstrated to pose increased risk for thrombolysis. Experience with unruptured arteriovenousmalformations and ruptured aneurysms that have been secured is very limited.

Lumbar puncture within 7 days is not regarded as an absolute contraindication although there are limited data on this scenario.

Direct oral anticoagulants and thrombolysis:

There are currently limited data on the safety of intravenous thrombolysis in patients taking direct oral anticoagulants. If the patient isknown to have not taken their anticoagulant within 48 hours and they have normal renal function then thrombolysis should be no greaterrisk than in unanticoagulated patients. When anticoagulation has been taken within 48 hours or this is unknown the options are 1) empiricreversal of the anticoagulant with a specific reversal agent (eg idarucizumab for dabigatran, andexanet alfa for rivaroxaban or apixaban)and then thrombolysis, 2) coagulation testing using the assay appropriate to the particular medication - calibrated factor Xa assay forapixaban or rivaroxaban, dilute thrombin time for dabigatran with subsequent thrombolysis if the level is deemed sufficiently low tojustify the risk or 3) immediate endovascular thrombectomy without thrombolysis if rapidly available and there is a suitable large vesselocclusion. The likely relative delay to obtain blood test results versus commence endovascular thrombectomy should be considered."Safe" levels of direct anticoagulants have not been established. Most consensus recommendations are based on trough levels observed inthe pivotal trials which is probably more conservative than the INR>1.7 criterion used for warfarin. Examples of suggested drug levelsthat may allow thrombolysis when a specific reversal agent is not available are dabigatran <40ng/mL, apixaban <10ng/mL and rivaroxaban<100ng/mL but these may evolve and careful individual risk benefit consideration is advised.

Key Info

Benefits and harms Substantial net benefits of the recommended alternative


42 of 157

Rationale

High-quality evidence suggests that the benefits of intravenous alteplase outweigh its harms if given within 4.5 hours in patients satisfyingspecific criteria. However, benefits have not been established beyond 4.5 h.

Lower dose alteplase (0.6mg/kg) did not meet non-inferiority criteria and therefore standard (0.9mg/kg) dose is recommended.

Alteplase significantly improved the overall odds of a good stroke outcome at 90 days when delivered within 4·5 h of stroke onset,based on meta-analyses of randomized controlled trials. Earlier treatment was associated with greater benefits. There was nosignificant benefit when alteplase was delivered after 4.5 h using standard clinical and non-contrast CT eligibility criteria. There wasan increased risk of fatal intracranial haemorrhage during the first few days in alteplase treated patients which was offset by laterdeaths in the control group such that there was a neutral effect on mortality at 90 days.

In one large randomised controlled trial with mostly Asian patients (N=3206), low-dose (0.6mg/kg) intravenous alteplase reduced therisk of intracerebral haemorrhage, but patients tended to have increased disability and the trial did not meet non-inferiority criteriacompared to 0.9mg/kg.

Quality of evidence

The overall quality of evidence is high, based on meta-analyses of large RCTs with low risk of bias.

High


For most patients the benefits in reduced disability would be preferred to the small risk of symptomatic hemorrhage.



For an Australian setting, a decision analytic model used outcome and cost information from patients with ischaemic stroke treatedwith tPA at a single hospital, and published clinical trial data for the no tPA group, to estimate cost-effectiveness over 12 months [28].Treatment with tPA within 4.5 hours was shown to be cost-effective compared to no tPA treatment (produces health gains for anacceptable additional cost to the alternative) at AU$2377 per life-year saved and AU$1478 per QALY [28]. International literaturehas also shown that treatment with tPA within 4.5 hours was either cost-effective or dominant (more effective and cost-saving) in thelong-term [29] [30] [31] [32].



Population: Adults with acute stroke

Intervention: Intravenous alteplase

Comparator: Control

Summary

A Cochrane review by Wardlaw et al (2014) [20] included 27 RCTs of thrombolytic agents for treatment of ischaemic stroke. Inmost trials, treatment began up to 6 hours after stroke. Death or dependency by the end of follow-up was significantly reduced inthe 10 trials using intravenous alteplase (OR 0.74, 95% CI 0.77 to 0.93), although there was significant heterogeneity. A strongereffect was seen when analysing intravenous alteplase given with 3 hours of stroke (OR 0.65, 95% CI 0.54 to 0.80) with nosignificant heterogeneity. However, intravenous alteplase was also associated with significant increases in 7 to 10-day mortality,driven largely by a substantially increased risk of fatal intracranial haemorrhage (OR 4.18, 95% CI 2.99 to 5.84). Despite theincrease in ICH in the short term, there was strong evidence for a net benefit of rt-PA treatment for death and dependency,particularly for rt-PA administered within 3 hours.


43 of 157

The benefits also appear to continue into the long term, although data is more limited. The IST-3 collaborative group(2013) [22] reported 18-month follow-up outcomes from an RCT (N = 2348) administering intravenous alteplase within 6 hours.Alteplase treatment was associated with increased number of patients alive and independent at 18 months (Oxford HandicapScale score 0-2, OR 1.28, 95% CI 1.03 to 1.57), although the increase in patients alive and with a favourable outcome was only ofborderline significance (OHS score 0-1, OR 1.23, 95% 0.98 to 1.55). There was no significant difference in death by 18 months.Given that alteplase was given within 6 hours in this study rather than within 3 hours, it is unclear whether early rt-PAadministration also leads to greater improvements in long-term outcomes.

OutcomeTimeframe



Control Intravenous alteplase

Certainty ineffect


Summary


the end offollow-up

End of follow-up

9 Critical

Odds Ratio 0.84(CI 95% 0.77 - 0.93)


(Randomized controlled)Follow up ranges from 1

week to >1 year

583per 1000

540per 1000


ModerateDue to seriousinconsistency

Intravenous alteplasereduces death or

dependency at the end offollow-up

Death7 to 10 days

9 Critical

Odds Ratio 1.44(CI 95% 1.18 - 1.76)


(Randomized controlled)Follow up 7 to 10 days

64per 1000

90per 1000


HighIntravenous alteplaseincreases death early

after stroke

Death at the endof follow-up

End of follow-up

Odds Ratio 1.06(CI 95% 0.94 - 1.2)


(Randomized controlled)Follow up ranges from 1

week to >1 year

185per 1000

194per 1000


HighThere is little or no

difference in death at theend of follow-up

Fatalintracranial

haemorrhage7-10 days

9 Critical

Odds Ratio 4.18(CI 95% 2.99 - 5.84)



6per 1000

25per 1000


High

Intravenous alteplaseincreases fatal

intracranial haemorrhageearly after stroke


44 of 157

Symptomaticintracranial

haemorrhage7-10 days

8 Critical

Odds Ratio 3.72(CI 95% 2.98 - 4.64)



18per 1000

64per 1000


High

Intravenous alteplaseincreases symptomatic

intracranial haemorrhageearly after stroke


Death ordependency at the

end of follow-up

Intervention reference:Systematic review [20]with included studies: Mori1992, NINDS 1995, ECASS1995, ECASS II 1998,ATLANTIS B 1999,ATLANTIS A 2000, Wang2003, ECASS 3 2008,EPITHET 2008, IST3 2012,Baseline/comparatorreference: Control arm ofreference used forintervention

Risk of bias: No seriousInconsistency: Serious The magnitude of statistical heterogeneity was high, with I^2:63%. ;Indirectness: No seriousImprecision: No seriousPublication bias: No serious

Death

Intervention reference:Systematic review [20]with included studies:Haley 1993, Mori 1992,ECASS 3 2008, Wang2003, ECASS II 1998,ECASS 1995, IST3 2012,EPITHET 2008, Haley1993, Mori 1992, ECASS 32008, Wang 2003, ECASSII 1998, ECASS 1995, IST32012, EPITHET 2008,Baseline/comparatorreference: Control arm ofreference used forintervention


Death at the end offollow-up

Intervention reference:Systematic review [20]with included studies:JTSG 1993, NINDS 1995,Mori 1992, Haley 1993,ECASS 1995, ECASS II1998, ATLANTIS B 1999,ATLANTIS A 2000, Wang2003, EPITHET 2008,ECASS 3 2008, IST3 2012,Baseline/comparatorreference: Control arm of

Risk of bias: No seriousInconsistency: No seriousIndirectness: No seriousImprecision: No serious 95%CI crosses 1 but it's unlikely to change clinical decision ;Publication bias: No serious


45 of 157

References

[20] Wardlaw JM Thrombolysis for acute ischaemic stroke. Cochrane Database of Systematic Review 2014; Pubmed


Fatal intracranialhaemorrhage

Intervention reference:Systematic review [20]with included studies:Haley 1993, NINDS 1995,ECASS 1995, ECASS II1998, ATLANTIS B 1999,ATLANTIS A 2000, ECASS3 2008, IST3 2012, Haley1993, NINDS 1995, ECASS1995, ECASS II 1998,ATLANTIS B 1999,ATLANTIS A 2000, ECASS3 2008, IST3 2012,Baseline/comparatorreference: Control arm ofreference used forintervention



haemorrhage

Intervention reference:Systematic review [20]with included studies:JTSG 1993, NINDS 1995,Mori 1992, Haley 1993,ECASS 1995, ECASS II1998, ATLANTIS B 1999,ATLANTIS A 2000, Wang2003, EPITHET 2008,ECASS 3 2008, IST3 2012,Baseline/comparatorreference: Control arm ofreference used forintervention



Population: Adults with acute stroke treated within 3 hours


Comparator: Control


46 of 157


Summary

An individual patient data meta-analysis conducted by Emberson et al (2014) [21] included subgroup analyses for patients treated<= 3 hours after stroke, > 3 and <= 4.5 hours, and > 4.5 hours. It showed that treatment with 3 hours was associated with thegreatest improvement on good stroke outcomes (mRS of 0 or 1) at 90 days (114 per 1000). Alteplase was associatedwith increased risk of intracranial haemorrhage within 7 days, and this effect appeared to greater for patients given alteplasewithin 3 hours (29 per 1000). Taken together, therefore, despite an average absolute increased risk of early death fromintracranial haemorrhage of about 2%, by 3–6 months this risk was off set by an average absolute increase in disability-freesurvival of about 10% for patients treated within 3·0 h.

A Cochrane review by Wardlaw et al (2014) [20] included 27 RCTs of thrombolytic agents for treatment of ischaemic stroke. Inmost trials, treatment began up to 6 hours after stroke. Death or dependency by the end of follow-up was significantly reduced inthe 10 trials using intravenous alteplase (OR 0.74, 95% CI 0.77 to 0.93), although there was significant heterogeneity. A strongereffect was seen when analysing intravenous alteplase given with 3 hours of stroke (OR 0.65, 95% CI 0.54 to 0.80) with nosignificant heterogeneity

OutcomeTimeframe




Certainty ineffect


Summary

Favourableoutcome

3-6 months

8 Critical

Odds Ratio 1.75(CI 95% 1.35 - 2.27)


(Randomized controlled)Follow up 3 to 6 months

231per 1000

345per 1000


HighIntravenous alteplase

within 3 hours increasesfavourable outcome

Death90 days

9 Critical

Hazard Ratio 1(CI 95% 0.81 - 1.24)



HighIntravenous alteplase

within 3 hours has littleor no difference on death

Fatalintracranial

haemorrhage7 days

9 Critical

Odds Ratio 10.86(CI 95% 2.54 - 46.41)



3per 1000

32per 1000


High

Intravenous alteplasewithin 3 hours increases

fatal intracranialhaemorrhage


Favourableoutcome

Intervention reference:Systematic reviewBaseline/comparatorreference: Control arm of

Risk of bias: No seriousInconsistency: No serious no heterogeneity analysis was done ;Indirectness: No serious


47 of 157

References


[21] Emberson J, Lees KR, Lyden P, Blackwell L, Albers G, Bluhmki E, Brott T, Cohen G, Davis S, Donnan G, Grotta J, Howard G,Kaste M, Koga M, von Kummer R, Lansberg M, Lindley RI, Murray G, Olivot JM, Parsons M, Tilley B, Toni D, Toyoda K, Wahlgren N,Wardlaw J, Whiteley W, del Zoppo GJ, Baigent C, Sandercock P, Hacke W Effect of treatment delay, age, and stroke severity on theeffects of intravenous thrombolysis with alteplase for acute ischaemic stroke: a meta-analysis of individual patient data fromrandomised trials.. Lancet (London, England) 2014;384(9958):1929-35- Pubmed Journal



Death


Risk of bias: No seriousInconsistency: No serious no heterogeneity analysis was done ;Indirectness: No seriousImprecision: No seriousPublication bias: No serious





Population: Adults with acute stroke treated at 3-4.5 hours


Comparator: Control

Summary

An individual patient data meta-analysis conducted by Emberson et al (2014) [21] included subgroup analyses for patients treated<= 3 hours after stroke, > 3 and <= 4.5 hours, and > 4.5 hours. It showed that treatment with 4.5 hours was associated with theimprovement on good stroke outcomes (mRS of 0 or 1) at 90 days (51 per 1000). Alteplase was associated with increased risk ofintracranial haemorrhage within 7 days (23 per 1000). Taken together, therefore, despite an average absolute increased risk ofearly death from intracranial haemorrhage of about 2%, by 3–6 months this risk was off set by an average absolute increase indisability-free survival of about 5% for patients treated after 3·0 h, up to 4·5 h.

OutcomeTimeframe



Control Intravenous alteplaseCertainty in

effectSummary


48 of 157



http://dx.doi.org/10.1016/S0140-6736(14)60584-5


Death90 days

9 Critical

Hazard Ratio 1.14(CI 95% 0.95 - 1.36)



High

Intravenous alteplasebetween 3 and 4.5 hourshas little or no difference

on death

Favourableoutcome

3-6 months

8 Critical

Odds Ratio 1.26(CI 95% 1.05 - 1.51)



301per 1000

352per 1000


High

Intravenous alteplasebetween 3 and 4.5 hours

increases favourableoutcome

Fatalintracranial

haemorrhage7 days

9 Critical

Odds Ratio 5.63(CI 95% 2.49 - 12.76)



5per 1000

28per 1000


High

Intravenous alteplasebetween 3 and 4.5 hours

increases fatalintracranial haemorrhage


Death



Favourableoutcome







49 of 157

References



Population: Adults with acute stroke treated at 4.5-6 hours


Comparator: Control

Summary

An individual patient data meta-analysis conducted by Emberson et al (2014) [21] included subgroup analyses for patients treated<= 3 hours after stroke, > 3 and <= 4.5 hours, and 4.5 - 6 hours. It showed that treatment between 4.5 -6 hours was associatedwith a small improvement on good stroke outcomes (mRS of 0 or 1) at 90 days (30 per 1000), but also increased risk offatal intracranial haemorrhage within 7 days (21 per 1000). In this case, it is unclear that if the benefits outweigh the potentialharms.

OutcomeTimeframe




Certainty ineffect


Summary

Death90 days

9 Critical

Hazard Ratio 1.22(CI 95% 0.99 - 1.5)



CI 95% HighIntravenous alteplase

after 4.5 hours slightlyincreases death

Favourableoutcome

3-6 months

8 Critical

Odds Ratio 1.15(CI 95% 0.95 - 1.4)



306per 1000

336per 1000


High

Intravenous alteplaseafter 4.5 hours slightly

increases favourableoutcome


50 of 157


http://dx.doi.org/10.1016/S0140-6736(14)60584-5

References


Fatalintracranial

haemorrhage7 days

9 Critical

Odds Ratio 8.16(CI 95% 2.88 - 23.11)



3per 1000

24per 1000


High

Intravenous alteplaseafter 4.5 hours increases

fatal intracranialhaemorrhage


Death



Favourableoutcome








Intervention: Low-dose intravenous alteplase

Comparator: Standard-dose intravenous alteplase


51 of 157


http://dx.doi.org/10.1016/S0140-6736(14)60584-5

Summary

Anderson et al (2016) [23] compared low-dose (0.6 mg per kilogram body weight) intravenous alteplase to standard dose (0.9 mgper kilogram) in an open-label randomised trial (N = 3310). While previous evidence on intravenous alteplase has suggested thatdoses of 0.9 mg per kilogram body weight provide benefits in the form of increased survival without disability, the treatment hasalso been associated with increased intracerebral haemorrhage, particularly in the short term. This risk of intracerebralhaemorrhage may be higher in Asian populations. A previous uncontrolled study showed similar benefits from low-dose alteplasewith decreased intracerebral haemorrhage. In this trial, low-dose alteplase was not shown to be non-inferior to standard dosetreatment when comparing the primary outcome of modified Rankin scale scores 2-6 (OR 1.09, 95% CI 0.95 to 1.25), where theboundary for non-inferiority was prespecified at 1.14. However, there were significantly fewer symptomatic intracerebralhaemorrhages in patients treated with low-dose alteplase (1% for the low-dose group vs 2.1% for the standard dose). The trialincluded predominantly Asian patients which could limit generalisability, but in subgroup analyses no significant differences wereseen between Asian and non-Asian patients.

Previous comparisons of dosages, included in a 2013 Cochrane review by Wardlaw et al [24], provided limited evidence on overallmortality or death and dependency. Only a few small trials reporting these outcomes were included in the review, with resultsfrom 5 studies (N =496) showing a lower number of total deaths in patients given higher-dose alteplase (OR 0.74, 95% 0.37 to1.52) but no significant differences. 4 included trials also showed significantly increased fatal intracranial haemorrhage but thetotal number of events was low, with 3 out of 4 included trials observing no fatal ICH.

OutcomeTimeframe



Standard-doseintravenous alteplase

Low-doseintravenous alteplase

Certainty ineffect


Summary

Death ordisability

90 days

9 Critical

Odds Ratio 1.09(CI 95% 0.95 - 1.25)



511per 1000

533per 1000


LowDue to serious

imprecision, Dueto serious

indirectness

Low-dose intravenousalteplase may slightly

increase death ordisability

Death90 days

9 Critical

Odds Ratio 0.8(CI 95% 0.63 - 1.01)



103per 1000

84per 1000


LowDue to serious


indirectness

Low-dose intravenousalteplase may slightly

decrease death

Improvedfunctionaloutcome

90 days

8 Critical

Odds Ratio 1(CI 95% 0.89 - 1.13)



LowDue to serious


indirectness

Low-dose intravenousalteplase may have little

or no difference onimproved functional

outcome


52 of 157

References


[23] Anderson CS, Robinson T, Lindley RI, Arima H, Lavados PM, Lee T-H, Broderick JP, Chen X, Chen G, Sharma VK, Kim JS, ThangNH, Cao Y, Parsons MW, Levi C, Huang Y, Olavarría VV, Demchuk AM, Bath PM, Donnan GA, Martins S, Pontes-Neto OM, Silva F,Ricci S, Roffe C, Pandian J, Billot L, Woodward M, Li Q, Wang X, Wang J, Chalmers J Low-Dose versus Standard-Dose IntravenousAlteplase in Acute Ischemic Stroke. N Engl J Med 2016; Journal Website

SymptomaticICH

90 days

8 Critical

Odds Ratio 0.48(CI 95% 0.27 - 0.86)



21per 1000

10per 1000


LowDue to serious


indirectness

Low-dose intravenousalteplase may decrease

symptomatic ICH


Death or disability


Risk of bias: No seriousInconsistency: No seriousIndirectness: Serious Mostly Asian population ;Imprecision: Serious Only data from one study ;Publication bias: No serious

Death



Improvedfunctional outcome



Symptomatic ICH




53 of 157


http://dx.doi.org/10.1016/S0140-6736(14)60584-5

http://dx.doi.org/10.1056/NEJMoa1515510


Practice Statement

Thrombolysis should be undertaken in a setting with appropriate infrastructure, facilities and network support (e.g. viatelemedicine) including:

• access to an interdisciplinary acute care team with expert knowledge of stroke management who are trained in delivery of

thrombolysis and monitoring of patients receiving thrombolytic therapy

• a streamlined acute stroke assessment workflow (including ambulance prenotification, code stroke team response and direct

transport from triage to CT scan) to minimise treatment delays and protocols available to guide medical, nursing and allied health

acute phase management

• immediate access to imaging facilities and staff trained to interpret images

• routine data collected in a central register to allow monitoring, benchmarking and improvements of patient outcomes over time

for those treated with reperfusion

• Where possible the patient and caregivers should be involved in the decision to give thrombolysis. However, as an emergency

therapy, formal consent for thrombolysis is not required.

4.2 - Neurointervention

From 2010 guideline:

Neurointerventional therapy in this section includes intra-arterial (IA) thrombolysis and mechanical clot removal. Most of the reportedstudies reported have been small observational studies based in highly specialised centres (i.e. those with advanced imaging, neurosurgicalspecialisation and appropriate infrastructure). There are currently only a few very large urban centres which offer such services in Australia.

One updated Cochrane review 12 identified four IA thrombolysis RCTs, two using urokinase and two using recombinant pro-urokinase. IAthrombolysis resulted in a significant reduction in the combined odds of death or dependency at follow-up (OR 0.49, 95% CI 0.31–0.79). Thelargest RCT (n=180) of IA thrombolysis (in addition to heparin) found high recanalisation rates (66% vs 18%, p<0.001), similar mortality,

improved outcomes (p=0.04) but higher sICH (10% vs 2%, p=0.06).235

One systematic review involving 13 case series describing outcomes for IA thrombolysis or IV thrombolysis in basilar artery occlusion found

no difference in outcomes between IV or IA thrombolysis even though significantly greater recanalisation rates were observed.236 One smallRCT of IA thrombolysis within 24 hours in posterior circulation strokes reported possible improved outcome but numbers were too small for

meaningful analysis.237 A subsequent large registry cohort found 68% of patients with BAO had a poor outcome (mRS >3). No statisticallysignificant superiority was found for any intervention strategy (antiplatelet therapy, IA or IV thrombolysis).238 More robust data arerequired.

No RCTs were found for mechanical clot removal. A systematic review (23 small observational studies) found those treated with clot removaldevices were 14.9 times (95% CI, 4.4–50.0) more likely to have a good outcome (mRS ≤ 2) and 2.2 times less likely to die (95% CI, 0.98–5.1)after adjustment for age, sex, and pre-intervention NIHSS.239 A good outcome (mRS<2) was reported in 36% of the pooled population.

However, the mortality rate was 29% and the haemorrhage rate 22%.239 More robust data are required.

Large tertiary centres could organise facilities for IA thrombolysis to allow appropriate referral for highly selected patients within themetropolitan areas. In rural areas, appropriate network arrangements would also facilitate such referral (see 1.2.5 Telemedicine andnetworks). It is highly likely that these services will continue to be limited due to the small number of trained personnel and associated highcosts.


54 of 157


Endovascular thrombectomy commenced within six hours of stroke onset should be undertaken in patients with ischaemic stroke causedby a large vessel occlusion in the internal carotid artery, proximal cerebral artery (M1 segment), or in tandem occlusion of both thecervical carotid and intracranial arteries.

Practical Info

(NOTE: list eligiblity criteria here)

Key Info

Rationale

Endovascular thrombectomy is effective in a broad range of patients without evidence of an effect of age, sex or clinical severity ontreatment benefit. Trials conducted have commenced thrombectomy procedure within 6 hours where there has been confirmed largeartery occlusion.

Benefits and harms

There is clear and high quality evidence of improved functional outcome and no evidence of increased risk of symptomaticintracerebral hemorrhage.


Quality of evidence

Five independent randomised trials from different health care settings produced highly consistent results and have been combined inan individual patient data meta-analysis.

High


Functional outcome benefits are clinically significant and important to patients and carers.



Three international economic evaluations showed that mechanical thrombectomy, predominantly performed with stent retrievers,combined with IV tPA was more effective and cost-saving i.e. dominant [43][45] or cost-effective [44] when compared to IV tPAalone.This finding was independent of regional differences in costs and how mechanical thrombectomy was performed.




Intervention: Endovascular mechanical thrombectomy

Comparator: Standard medical care

Summary

Goyal et al. (2016) [38] conducted an individual patient meta-analysis that pooled results from five recent trials of endovascularthrombectomy. The included trials were different to previously published trials in that they all used CT or magnetic resonanceimaging to target large vessel occlusions, emphasised fast treatment, and used second-generation neurothrombectomy devices


55 of 157

with better recanalisation rates and lower complications. The primary outcome was score on the modified Rankin scale, analysedusing ordinal logistic regression to estimate the odds that intervention would improve mRS scores by 1 or more points.Intervention was shown to increase the odds of improvement significantly (common odds ratio 2.49, 95% CI 1.76 to 3.53). Thenumber needed to treat for one patient to have a reduction of their mRS score of 1 point or more was 2.6. The dichotomousoutcome of mRS 0-2 vs 3-6 also showed a significant increase in functional independence (adjusted OR 2.71, 95% CI 2.07 to 3.55).There were no significant effects on 90-day mortality or symptomatic intracranial haemorrhage. The trials were generally of highquality, with blinded outcome assessment, and effects were consistent across trials.

A previous meta-analysis by Badhiwala et al. (2015) [35] included the same 5 trials as the Goyal et al. analysis but also included 3earlier trials. The pooled results showed a significant increase in the odds of a reduction of mRS score (OR 1.56, 95% CI 1.14 to2.13), and in the odds of functional independence (OR 1.71, 95% CI 1.18 to 2.49), although in both cases the effect appeared to beweaker than the comparable analysis in the Goyal et al. meta-analysis.

The stronger effects in the Goyal et al. analysis may result from the newer trials employing different treatment methods asdiscussed above. This appears to be supported by a significant interaction between functional outcome and year of publication inthe Badhiwala et al. analysis. This further informs the best practice in endovascular mechanical thrombectomy.

OutcomeTimeframe



Standard medicalcare

Endovascularmechanical

thrombectomy

Certainty ineffect


Summary

Improvedfunctionaloutcome3 months

8 Critical

Odds Ratio 2.49(CI 95% 1.76 - 3.53)



Difference: 385 moreCI 95%

HighEndovascular

stentthrombectomy

reduces disabilitywith high

confidence basedon 5 high quality

randomizedcontrolled trialswith consistent

effects.

Endovascular mechanicalthrombectomy improves

functional outcome


3 months

8 Critical

Odds Ratio 2.71(CI 95% 2.07 - 3.55)



265per 1000

494per 1000


HighEndovascular

stentthrombectomy



randomizedcontrolled trialswith consistent

effects.

Endovascular mechanicalthrombectomy improvesfunctional independence


56 of 157

Mortality3 months

9 Critical

Odds Ratio 0.73(CI 95% 0.47 - 1.13)



189per 1000

145per 1000

Difference: 44 fewer per 1000( CI 95% 19 more - 90 fewer )

HighOverall mortalitydid not differ with

endovascularmechanical

thrombectomy.However the

subgroup aged>80 years did have

a significantreduction in

mortality.

Endovascular mechanicalthrombectomy has littleor no effect on mortalityoverall with a significant

reduction in patientsaged > 80 years


haemorrhagewithin 36 hours of

treatment

8 Critical

Odds Ratio 1.07(CI 95% 0.62 - 1.84)



43per 1000

46per 1000

Difference: 3 more per 1000( CI 95% 33 more - 16 fewer )

HighNo difference in

the rate ofsymptomatichaemorrhage

Endovascular mechanicalthrombectomy has little

or no difference onsymptomatic intracranial

haemorrhage




Risk of bias: No serious Inability to blind participants. Assessors blinded. Low risk ofpotential performance bias. Some trials stopped earlier than scheduled, resulting in potentialfor overestimating benefits ;Inconsistency: No seriousIndirectness: No seriousImprecision: No seriousPublication bias: No serious One of five trials commercially sponsored, partial funding ofother trials by commercial unrestricted grants. ;




Mortality




57 of 157


Endovascular thrombectomy should commence as early as possible. If eligible, patients should receive intravenous alteplase whileconcurrently arranging endovascular thrombectomy with neither treatment delaying the other.

Key Info

References

[35] Badhiwala JH, Nassiri F, Alhazzani W, Selim MH, Farrokhyar F, Spears J, Kulkarni AV, Singh S, Alqahtani A, Rochwerg B,Alshahrani M, Murty NK, Alhazzani A, Yarascavitch B, Reddy K, Zaidat OO, Almenawer SA Endovascular Thrombectomy for AcuteIschemic Stroke: A Meta-analysis.. JAMA 2015;314(17):1832-43- Pubmed Journal

[38] Goyal M, Menon BK, van Zwam WH, Dippel DWJ, Mitchell PJ, Demchuk AM, Dávalos A, Majoie CBLM, van der Lugt A, deMiquel MA, Donnan GA, Roos YBWEM, Bonafe A, Jahan R, Diener H-C, van den Berg LA, Levy EI, Berkhemer OA, Pereira VM,Rempel J, Millán M, Davis SM, Roy D, Thornton J, Román LS, Ribó M, Beumer D, Stouch B, Brown S, Campbell BCV, vanOostenbrugge RJ, Saver JL, Hill MD, Jovin TG Endovascular thrombectomy after large-vessel ischaemic stroke: a meta-analysis ofindividual patient data from five randomised trials.. Lancet (London, England) 2016;387(10029):1723-31- Pubmed Journal


haemorrhage



Benefits and harms

There is clear and high quality evidence of improved functional outcome and no evidence of increased risk of symptomaticintracerebral hemorrhage.


Quality of evidence

Five independent randomised trials from different health care settings produced highly consistent results and have been combined inan individual patient data meta-analysis.

High




Resources and other considerations No important issues with the recommended alternative


58 of 157


http://dx.doi.org/10.1001/jama.2015.13767


http://dx.doi.org/10.1016/S0140-6736(16)00163-X

Rationale

As with IV thrombolysis, time is brain and earlier removal of occlusion is more likely to lead to improved outcomes. Trials to date haveusually included both IV tPA and clot retrieval. However, endovascular thrombectomy is effective in patients with contraindications tointravenous thrombolysis.

Three international economic evaluations showed that mechanical thrombectomy, predominantly performed with stent retrievers,combined with IV tPA was more effective and cost-saving i.e. dominant [43][45] or cost-effective [44] when compared to IV tPA alone.This finding was independent of regional differences in costs and how mechanical thrombectomy was performed.



Intervention: Endovascular mechanical thrombectomy + intravenous thrombolysis

Comparator: Intravenous thrombolysis alone

Summary

Goyal et al (2016) [38] conducted an individual patient meta-analysis that pooled results from five recent trials of endovascularthrombectomy. The overall analysis showed a significant increase in odds of a reduced modified Rankin scale score (OR 2.49, 95%CI 1.76 to 3.53). A prespecified subgroup analysis of patients who had received alteplase treatment found a similar treatmenteffect (OR 2.45, 95% CI 1.68 to 3.57). There was non-significant heterogeneity (p = 0.43) between subgroups receiving or notreceiving alteplase, suggesting that the effect of thrombectomy did not differ between the groups. As in the overall analysis,endovascular thrombectomy significantly improved the odds of functional independence and produced no significant differencesin 90-day mortality.

Palesch et al (2015) [33] reported 12 month outcomes from an earlier trial of endovascular therapy (IMS III), where all patients (inboth the endovascular therapy and control groups) had received intravenous alteplase. At 12 months, the odds of functionalindependence following endovascular therapy were significantly improved for patients with severe strokes but showed nodifference among patients with moderate stroke. However, the more recent trials included in the Goyal et al. analysis all includeda number of improved treatment methods, and short term treatment outcomes following endovascular thrombectomy appear tobe substantially improved in these newer trials. It is unclear whether the same improvements occur for longer term outcomes.

OutcomeTimeframe



Intravenousthrombolysis alone

Endovascularmechanical

thrombectomy +intravenous

thrombolysis

Certainty ineffect


Summary


3 months

8 Critical

Odds Ratio 1.67(CI 95% 1.37 - 2.05)



270per 1000

382per 1000


HighEndovascular

stentthrombectomy



randomizedcontrolled trials

Endovascular mechanicalthrombectomy improvesfunctional independence


59 of 157

with consistenteffects.

Mortality3 months

9 Critical

Odds Ratio 0.75(CI 95% 0.5 - 1.12)



184per 1000

145per 1000


HighOverall mortalitydid not differ with

endovascularmechanical

thrombectomy.However the

subgroup aged>80 years did have

a significantreduction in

mortality.

Endovascular mechanicalthrombectomy has littleor no effect on mortalityoverall with a significant

reduction in patientsaged > 80 years


8 Critical

Odds Ratio 2.45(CI 95% 1.68 - 3.57)



CI 95%

HighEndovascular

thrombectomy inaddition to

alteplase improvesfunctionaloutcome

Endovascular mechanicalthrombectomy +

intravenous thrombolysisimproves functional

outcome





Mortality





Risk of bias: No serious Inability to blind participants. Assessors blinded. Low risk ofpotential performance bias. Some trials stopped earlier than scheduled, resulting in potentialfor overestimating benefits ;Inconsistency: No seriousIndirectness: No seriousImprecision: No serious


60 of 157

References


[40] Palesch YY, Yeatts SD, Tomsick TA, Foster LD, Demchuk AM, Khatri P, Hill MD, Jauch EC, Jovin TG, Yan B, von Kummer R,Molina CA, Goyal M, Schonewille WJ, Mazighi M, Engelter ST, Anderson C, Spilker J, Carrozzella J, Ryckborst KJ, Janis LS, SimpsonA, Simpson KN, Broderick JP Twelve-Month Clinical and Quality-of-Life Outcomes in the Interventional Management of Stroke IIITrial.. Stroke; a journal of cerebral circulation 2015;46(5):1321-7- Pubmed Journal

Publication bias: No serious One of five trials commercially sponsored, partial funding ofother trials by commercial unrestricted grants. ;


Population: Adults with stroke ineligible for IV thrombolysis


Comparator: Standard care

Summary

Goyal et al (2016) [38] conducted an individual patient meta-analysis that pooled results from five recent trials of endovascularthrombectomy. The overall analysis showed a significant increase in odds of a reduced modified Rankin scale score (OR 2.49, 95%CI 1.76 to 3.53). A prespecified subgroup analysis of patients ineligible for alteplase (N = 188) found a similar effect (OR 2.43, 95%CI 1.30 to 4.55), with non-significant heterogeneity (p = 0.43) between patients eligible and not eligible for alteplase.Endovascular thrombectomy appears to equally effective among patients eligible and ineligible for alteplase.

OutcomeTimeframe



Standard care Endovascularmechanical

thrombectomy

Certainty ineffect


Summary


8 Critical

Odds Ratio 2.43(CI 95% 1.3 - 4.55)



CI 95%

HighEndovascular

thrombectomyimproves outcome

in patientsineligible for

alteplase

Endovascular mechanicalthrombectomy improves

functional outcome inpatients ineligible for

intravenous thrombolysis


61 of 157


http://dx.doi.org/10.1016/S0140-6736(16)00163-X


http://dx.doi.org/10.1161/STROKEAHA.115.009180

References

[38] Goyal M, Menon BK, van Zwam WH, Dippel DWJ, Mitchell PJ, Demchuk AM, Dávalos A, Majoie CBLM, van der Lugt A, deMiquel MA, Donnan GA, Roos YBWEM, Bonafe A, Jahan R, Diener H-C, van den Berg LA, Levy EI, Berkhemer OA, Pereira VM,Rempel J, Millán M, Davis SM, Roy D, Thornton J, Román LS, Ribó M, Beumer D, Stouch B, Brown S, Campbell BCV, van


3 months

8 Critical

Odds Ratio 2.05(CI 95% 1.16 - 3.63)



223per 1000

370per 1000


HighRobust statistical

significance and noevidence of effect

heterogeneity.

endovascular mechanicalthrombectomy improvesfunctional independencein patients ineligible for

intravenous thrombolysis

Mortality3 months

9 Critical

Odds Ratio 1.11(CI 95% 0.6 - 2.07)



225per 1000

231per 1000


High

Endovascular mechanicalthrombectomy in

patients ineligible forintravenous thrombolysishas little or no difference

on mortality








Mortality


Risk of bias: No serious Inability to blind participants. Assessors blinded. Low risk ofpotential performance bias. Some trials stopped earlier than scheduled, resulting in potentialfor overestimating benefits. ;Inconsistency: No seriousIndirectness: No seriousImprecision: No seriousPublication bias: No serious One of five trials commercially sponsored, partial funding ofother trials by commercial unrestricted grants. ;


62 of 157


Endovascular thrombectomy shouldmay be performedconsidered in selected patients with occlusion of the basilar artery (given the poornatural history).

Needs rationale and any practical considerations

Practical Info

The evidence for benefit in ICA, M1 and tandem cervical ICA-M1 occlusions is clear cut. Cervical arterial dissection is potentially moretechnically challenging and not specifically studied in the randomized trials.M2 occlusions are relatively uncommon and under-represented in the randomized trials. Their anatomy and calibre is quite variable whichimpacts on the technical feasibility of the procedure. Observational and historically controlled data suggests there may be benefit in M2occlusions. At the present time decisions need to be made in the individual case based on the accessibility of clot and tissue at risk (clinicaldeficit and imaging pattern).Basilar artery occlusions are quite variable. Distal basilar occlusion is generally embolic in contrast to mid-basilar occlusion where there isoften an underlying atherosclerotic stenosis. A randomized trial of endovascular thrombectomy (BASICS) is ongoing but in the interimexpert opinion supports treatment of basilar artery occlusion based on the dire natural history and data indicating the benefit ofsuccessful recanalization. The time window for treatment of basilar artery occlusion may be longer than in the anterior circulation butthere is limited evidence to guide this.

Key Info

Oostenbrugge RJ, Saver JL, Hill MD, Jovin TG Endovascular thrombectomy after large-vessel ischaemic stroke: a meta-analysis ofindividual patient data from five randomised trials.. Lancet (London, England) 2016;387(10029):1723-31- Pubmed Journal

Benefits and harms

There is clear and high quality evidence of improved functional outcome and no evidence of increased risk of symptomaticintracerebral hemorrhage. Time window, very mild cases etc, dire prognosis untreated


Quality of evidence

Five independent randomised trials from different health care settings produced highly consistent results and have been combined inan individual patient data meta-analysis. One small randomized trial and registry studies have demonstrated major improvements inoutcome with successful recanalization. The registry study (BASICS) did not show clear differences between those treated withintravenous thrombolysis versus thrombectomy. A randomized trial is ongoing.

High






63 of 157


http://dx.doi.org/10.1016/S0140-6736(16)00163-X

Practice Statement

Endovascular thrombectomy may be considered in the following situations based on individual patient and advanced imaging factors:

• commencement of procedure surgery beyond 6within 12 hours

• occlusion in more distal middle cerebral artery (M2 segment)

• Endovascular thrombectomy should be performed by an experienced neurointerventionist with recognised training in the

procedure (Conjoint Committee for Recognition of Training in Interventional Neuroradiology).

Add Rationale if needed

Practical Info

Endovascular thrombectomy should be performed by an experienced neurointerventionist with recognised training in the procedure(Conjoint Committee for Recognition of Training in Interventional Neuroradiology). Refer to website: www.ccinr.org.au.

Key Info

Available evidence indicates that endovascular thrombectomy in the anterior circulation is highly cost effective (dominant inmodelling)

Benefits and harms

There is clear and high quality evidence of improved functional outcome and no evidence of increased risk of symptomaticintracerebral hemorrhage when the procedure is commenced 0-6 hours after stroke onset. Only 69 patients were treated beyond 6hours in the randomised trials and the benefit was not significant with those small numbers. Non-randomised data from theDEFUSE-2 study suggested benefit of reperfusion in patients with favorable perfusion-diffusion imaging up to 12 hours after onset.The POSITIVE, DAWN and DEFUSE 3 randomised trials are currently examining delayed window treatment using imaging selection.


Quality of evidence

Five independent randomised trials from different health care settings produced highly consistent results and have been combined inan individual patient data meta-analysis. Limited patient numbers treated beyond 6 hours limits confidence in interpretation in theextended time window.

High





Available evidence indicates that endovascular thrombectomy is highly cost effective (dominant in modelling)



64 of 157


Population: Adults with stroke onset > 6 hours



Summary

Goyal et al. (2016) [38] conducted an individual patient meta-analysis that pooled results from five recent trials of endovascularthrombectomy. The overall analysis showed a significant increase in odds of a reduced modified Rankin scale score (OR 2.49, 95%CI 1.76 to 3.53). In a prespecified subgroup analysis, data from patients who were randomised > 300 minutes after stroke onsetwere analysed. Generally patients were randomised < 420 minutes after onset, corresponding to beginning treatment less than 8hours after symptom onset. Patients still showed significant benefit from endovascular thrombectomy treatment (OR 1.76, 95%CI 1.05 to 2.97), although the benefit appeared smaller and the test of heterogeneity comparing patients randomised <= 300minute to > 300 minutes was near significance (p = 0.10). Unlike in the overall analysis, the odds of functional independence werenot significantly different for patients treated with endovascular thrombectomy, although the effect estimate was still in thedirection of benefit. There were no differences in 90-day mortality. Endovascular thrombectomy appears to still be effective > 6hours after stroke onset but the benefits may be smaller.

OutcomeTimeframe



Standard care Endovascularmechanical

thrombectomy

Certainty ineffect


Summary


8 Critical

Odds Ratio 1.76(CI 95% 1.05 - 2.97)



CI 95%


imprecision

Endovascular mechanicalthrombectomy beyond 6hours probably improves

functional outcome inselected patients


3 months

8 Critical

Odds Ratio 1.55(CI 95% 0.93 - 2.58)



262per 1000

355per 1000



imprecision

Endovascular mechanicalthrombectomy beyond 6hours probably improvesfunctional independence

in selected patients

Mortality3 months

9 Critical

Odds Ratio 1.15(CI 95% 0.53 - 2.5)



126per 1000

142per 1000



imprecision

Endovascular mechanicalthrombectomy beyond 6hours probably has littleor no effect on mortality



65 of 157

References




Risk of bias: No serious Inability to blind participants. Assessors blinded. Low risk ofpotential performance bias. Some trials stopped earlier than scheduled, resulting in potentialfor overestimating benefits. ;Inconsistency: No seriousIndirectness: No seriousImprecision: Serious Low number of patients with most treated soon after 6 hours. ;Publication bias: No serious One of five trials commercially sponsored, partial funding ofother trials by commercial unrestricted grants. ;



Risk of bias: No serious Inability to blind participants. Assessors blinded. Low risk ofpotential performance bias. Some trials stopped earlier than scheduled, resulting in potentialfor overestimating benefits. ;Inconsistency: No seriousIndirectness: No seriousImprecision: Serious Low number of patients with most treated soon after 6 hours. ;Publication bias: No serious One of five trials commercially sponsored, partial funding ofother trials by commercial unrestricted grants. ;

Mortality


Risk of bias: No serious Inability to blind participants. Assessors blinded. Low risk ofpotential performance bias. Some trials stopped earlier than scheduled, resulting in potentialfor overestimating benefits. ;Inconsistency: No seriousIndirectness: No seriousImprecision: Serious Low number of patients, Wide confidence intervals ;Publication bias: No serious One of five trials commercially sponsored, partial funding ofother trials by commercial unrestricted grants. ;


66 of 157


http://dx.doi.org/10.1016/S0140-6736(16)00163-X

5 - Antithrombotic therapy


An updated Cochrane review (12 RCTs) found consistent but modest net reduction in death or disability (NNT=79) of acute phase antiplatelet

therapy in ischaemic stroke.240 Almost all of the data were from trials with aspirin therapy (160–300 mg) that commenced within 48 hours

and continued in the weeks following stroke onset. 240 While there was a small increase in intracranial haemorrhage there was a definite net

benefit. 240 Two RCTs testing early (within 24 hours) clopidogrel plus aspirin 241 or extended-release dipyridamole plus aspirin 242 in patientswith TIA or minor ischaemic stroke have reported similar or potential benefits compared to aspirin monotherapy. Large (well-powered) RCTsare needed before therapy other than aspirin alone can be recommended in routine clinical care.

If patients receive thrombolysis, aspirin should be deferred for at least 24 hours and only prescribed if follow-up brain imaging has excludedintracranial haemorrhage (see 4.1 Thrombolysis). Another updated Cochrane review (24 RCTs) found no evidence in ischaemic stroke thatearly anticoagulant therapy (standard unfractionated heparin, low-molecular-weight heparins, heparinoids, oral anticoagulants or thrombin

inhibitors) reduced the odds of death or dependency (OR 0.99, 95% CI 0.93–1.04).243 Although there were fewer recurrent ischaemic strokes(OR 0.76, 95% CI 0.65–0.88) and fewer pulmonary emboli (OR 0.60, 95% CI 0.44–0.81), Early anticoagulant therapy was associated with an

increase in symptomatic ICH (OR 2.55, 95% CI 1.95–3.33) and extra-cranial haemorrhages (OR 2.99, 95% CI 2.24–3.99).243 Another meta-analysis (seven RCTs) similarly found no overall effect of anticoagulant treatment on death or dependency in acute cardio-embolic stroke (OR

1.0, 95% CI 0.82–1.24).244

Uncommon presentations such as arterial dissection may prompt consideration of early anticoagulation. Arterial dissection involves a teardeveloping in the inner lining of the artery, which increases the likelihood of clotting and stroke. Dissection is rare (2.5% of all strokes) but is

more frequent (5–22%) for patients less than 45 years old.245 There is currently no RCT evidence for the preferred antithrombotic therapy. Asystematic review of 26 small, lower-level studies suggested there was no difference in outcomes between antiplatelet and anticoagulation

therapy with only a small number (0.5%) of ICH.245


Patients with ischaemic stroke who are not receiving reperfusion therapy should receive antiplatelet therapy as soon as possible once

brain imaging has excluded haemorrhage.


Key Info

Rationale

High-quality evidence shows that aspirin significantly reduces death and dependency and recurrent stroke, with a small increase inintracranial haemorrhage in patients of ischaemic stroke who are not receiving reperfusion therapy.

Benefits and harms

Aspirin was shown to have small but statistically significant benefit in outcomes of death (9 fewer per 1000), death and dependency(13 fewer per 1000), and recurrent stroke (7 fewer per 1000). It was shown to increase symptomatic intracranial haemorrhage butthe effect was small (2 more per 1000 patients).

Small net benefit, or little difference between alternatives

Quality of evidence

The evidence is based on large RCTs with low risk of bias, reporting consistent results.

High


Cost-utility modelling study in Australia showed that treatment with aspirin within 48 hours of ischaemic stroke was cost-effective (AU$1,847 per DALY saved) [55].



67 of 157


Population: Adults with acute ischaemic stroke

Intervention: Aspirin

Comparator: Placebo or no treatment

Summary

Sandercock et al (2014) [43] conducted a systematic review and meta-analysis of immediate oral antiplatelet therapy for acuteischaemic stroke. 8 randomised controlled trials with 41,483 patients were included. The two largest trials, contributing 98% ofthe data, used 300mg or 160mg aspirin. Aspirin was associated with a small but significant reduction in death or dependence (OR0.95, 95% CI 0.91 to 0.99) at the end of follow-up (up to 6 months). There were also significant reductions in death and recurrentstroke, as well as a significant increase in symptomatic intracranial haemorrhages that was small in absolute terms due to the lowoverall risk. The review authors rated the risk of bias as low. Although one of the large trials contributing the majority of the datawas unblinded, outcomes were self-reported by patients or assessed by a blinded interviewer, and a pilot study suggested that themajority of patients did not remember the treatment they had received at 6-month follow-up.

Rothwell et al (2014) [39] conducted an individual patient data analysis of the effects of aspirin on risk of recurrent strokefollowing TIA or ischaemic stroke. Data for aspirin following acute stroke predominantly came from the two largest trials includedin the Sandercock et al (2014) review. Time course analysis of the risk of recurrent ischaemic stroke following aspirin treatmentwas conducted. For patients with mild or moderately severe neurological deficits, there was a non-significant reduction in risk inthe first 24 hours following aspiring treatment, with significant reductions by day 2 that remained significant at day 3, days 4-6and days 7-14. Risks were not significantly different after 14 days.

OutcomeTimeframe



Placebo or notreatment

Aspirin

Certainty ineffect


Summary

Death ordependence

End of follow-up

9 Critical

Odds Ratio 0.95(CI 95% 0.91 - 0.99)Based on data from41,291 patients in 4


Follow up 3 to 6 months

462per 1000

449per 1000


HighAspirin decreases death

or dependence

DeathEnd of follow-up

9 Critical



Follow up 3 to 6 months

129per 1000

120per 1000


High Aspirin decreases death

Recurrentstroke

During treatment

8 Critical



31per 1000

24per 1000


HighAspirin decreasesrecurrent stroke


68 of 157

References

[46] Rothwell PM, Algra A, Chen Z, Diener H-C, Norrving BO, Mehta Z Effects of aspirin on risk and severity of early recurrentstroke after transient ischaemic attack and ischaemic stroke: time-course analysis of randomised trials. The Lancet 2016; JournalWebsite

[50] Sandercock PAG, Counsell C, Tseng M-C, Cecconi E Oral antiplatelet therapy for acute ischaemic stroke.. The Cochranedatabase of systematic reviews 2014;3 CD000029- Pubmed Journal

Follow up 5 days to 3months of treatment


haemorrhageDuring treatment

8 Critical

Odds Ratio 1.22(CI 95% 1 - 1.5)

Based on data from40,850 patients in 3


Follow up 5 days to 3months of treatment

8per 1000

10per 1000


HighAspirin slightly increasessymptomatic intracranial

haemorrhage


Death ordependence

Intervention reference:Systematic reviewBaseline/comparatorreference: Systematicreview


Death



Recurrent stroke




haemorrhage




69 of 157

http://dx.doi.org/10.1016/S0140-6736(16)30468-8

http://dx.doi.org/10.1016/S0140-6736(16)30468-8




Acute antiplatelet therapy should be deferred for 24 hours if alteplase has been given.


Key Info

Rationale

Based on moderate quality of evidence, concurrent use of antiplatelets with alteplase probably increases symptomatic intracranialhaemorrhage with no apparent benefits. Therefore, acute antiplatelet therapy should be deferred when alteplase is given.

Benefits and harms

Addition of intravenous aspirin to alteplase versus alteplase alone showed no improvements in favourable outcomes (defined asmodified Rankin Scale 0-2) and an increase in symptomatic intracranial haemorrhage (28 more per 1000).


Quality of evidence

The quality of evidence is moderate as it comes from only one study which terminated early before reaching the powered sample size.

Moderate


Population: Adult stroke patients treated with alteplase

Intervention: Early antiplatelet therapy

Comparator: No additional therapy

Summary

A randomised trial (Zinstok and Roos 2012 [47]) comparing addition of intravenous aspirin to alteplase versus alteplase alone washalted early due to increased numbers of symptomatic intracranial haemorrhage in the aspirin group, with no evidence of benefiton the primary endpoint of a favourable outcome (score of 0-2 on the modified Rankin Scale).

OutcomeTimeframe



No additionaltherapy

Early antiplatelettherapy

Certainty ineffect


Summary

Death3 months

9 Critical

n/a



97per 1000

112per 1000

CI 95%

LowDue to very

seriousimprecision

early antiplatelet therapy(within 24h) may increase

death


70 of 157

References

[54] Zinkstok SM, Roos YB Early administration of aspirin in patients treated with alteplase for acute ischaemic stroke: arandomised controlled trial.. Lancet (London, England) 2012;380(9843):731-7- Pubmed Journal

Favourableoutcome3 months

8 Critical

Odds Ratio 0.91(CI 95% 0.66 - 1.26)



572per 1000

549per 1000



imprecision

early antiplatelet therapy(within 24h) probably haslittle or no difference on

favourable outcome


haemorrhage

8 Critical

Relative risk 2.78(CI 95% 1.01 - 7.63)



16per 1000

44per 1000



imprecision

early antiplatelet therapy(within 24h) probably

increases symptomaticintracranial haemorrhage


Death


Risk of bias: No seriousInconsistency: No seriousIndirectness: No seriousImprecision: Very Serious The study was terminated early due to futility and didn't reach thepowered sample size N = 600, Only data from one study; no relative effect estimate ;Publication bias: No serious

Favourableoutcome


Risk of bias: No seriousInconsistency: No seriousIndirectness: No seriousImprecision: Serious The study was terminated early due to futility and didn't reach thepowered sample size N = 600., Only data from one study ;Publication bias: No serious


haemorrhage


Risk of bias: No seriousInconsistency: No seriousIndirectness: No seriousImprecision: Serious The study was terminated early due to futility and didn't reach thepowered sample size N = 600., Only data from one study ;Publication bias: No serious


71 of 157


http://dx.doi.org/10.1016/S0140-6736(12)60949-0


Routine use of early anticoagulation in unselected patients ( i . e . without c a r d i o e m b o l i s m ( e . g . AF) following TIA/stroke is not

recommended.


Key Info

Rationale

High-quality evidence suggests that anticoagulants did not show any benefits in death and dependency in patients with acute ischaemicstroke. Therefore, routine use of anticoagulation in patients with no indication of potential benefits from it is not recommended.

Benefits and harms

Anticoagulation was shown to significantly reduce recurrent stroke but to increase symptomatic intracranial haemorrhage of similarextent. These effects appeared to produce a non-significant difference in death and dependency at the follow-up of greater than amonth.


Quality of evidence

Quality of evidence is high - multiple large randomised controlled trials reporting consistent results.

High



Intervention: Anticoagulant

Comparator: Placebo or no treatment

Summary

A Cochrane review by Sandercock et al (2015) [40] assessed the effectiveness of early anticoagulation following acute ischaemicstroke. 24 trials with 23,748 participants were included. Meta-analysis of 8 trials reporting death or dependence with a follow-upgreater than 1 month showed no difference in the odds of death or dependency (OR 0.99, 95% CI 0.93 to 1.04). There wassubstantial heterogeneity in this analysis, with low-molecular-weight heparins and subcutaneous heparinoids showing non-significant benefit and direct thrombin inhibitors showing non-significant harms. While anticoagulants significantly decreasedrecurrent stroke during the treatment period, they also significantly increased symptomatic intracranial haemorrhage and thesetwo effects appeared to produce no difference in overall death or dependency.

Based on various international guidelines that recommend targeting of heparin treatment at stroke patients with high risk ofvenous thrombotic events or low risk of haemorrhagic events, Whiteley et al (2013) [46] conducted an individual patient datameta-analysis of the 5 largest randomised controlled trials of heparin treatment. They found no evidence that patients predictedto be at higher risk of thrombotic events or lower risk of haemorrhagic events benefited from treatment with heparins. Theysuggested that existing guidelines recommending targeted selection of patients for heparin treatment be revised.

OutcomeTimeframe


Absolute effect estimatesCertainty in

effectSummary


72 of 157

Placebo or notreatment

Anticoagulantestimates(Quality ofevidence)

Death ordependence

End of follow-up

9 Critical



Follow up >30 days

599per 1000

597per 1000



Anticoagulant probablyhas little or no differenceon death or dependence

DeathEnd of follow-up

9 Critical

Odds Ratio 1.05(CI 95% 0.98 - 1.12)Based on data from

22,776 patients in 11studies. (Randomized

controlled)Follow up > 30 days

205per 1000

213per 1000


Highanticoagulant has little or

no difference on death

Recurrentstroke

During treatment

8 Critical



controlled)Follow up 7 to 30 days of

treatment

36per 1000

28per 1000


HighAnticoagulant decreases

recurrent stroke


haemorrhageDuring treatment

8 Critical



controlled)Follow up 7 to 30 days of

treatment

5per 1000

13per 1000


HighAnticoagulant increases

symptomatic intracranialhaemorrhage


Death ordependence


Risk of bias: No seriousInconsistency: Serious Point estimates vary widely ;Indirectness: No seriousImprecision: No seriousPublication bias: No serious

Death




73 of 157

Weak Recommendation

Aspirin plus clopidogrel may should be used in the short term (first three weeks) in high risk patients with after minor ischaemic stroke or

TIA to prevent stroke recurrence.


Practical Info

Include details on loading doses here as well as current PBS criteria

Key Info

References

[47] Sandercock PAG, Counsell C, Kane EJ Anticoagulants for acute ischaemic stroke.. The Cochrane database of systematicreviews 2015;3 CD000024- Pubmed Journal

[53] Whiteley WN, Adams HP, Bath PMW, Berge E, Sandset PM, Dennis M, Murray GD, Wong K-SL, Sandercock PAG Targeted useof heparin, heparinoids, or low-molecular-weight heparin to improve outcome after acute ischaemic stroke: an individual patientdata meta-analysis of randomised controlled trials.. The Lancet. Neurology 2013;12(6):539-45- Pubmed Journal

Recurrent stroke




haemorrhage



Benefits and harms

A systematic review with the majority of data coming from the comparison of clopidogrel plus aspirin versus aspirin alone in trialCHANCE showed significant reduction in recurrent stroke and the composite outcome of stroke, TIA, ACS and all deaths, with noincrease in major bleeding.


Quality of evidence

Quality of evidence is moderate because the majority of population was Asian - the applicability to Australian population is uncertain.

Moderate


74 of 157




http://dx.doi.org/10.1016/S1474-4422(13)70079-6

Rationale

Evidence, predominantly from CHANCE trial, shows that the combination of aspirin and clopidogrel significantly improved patientsoutcomes without an increase in harms. However, as the majority of participants was Asian population, it is unclear if these conclusionscan be directly applied to stroke patients in Australia.



Intervention: Dual antiplatelet therapy

Comparator: Mono antiplatelet therapy

Summary

A systematic review (Wong et al 2013 [45]) of trials investigating dual antiplatelet therapy compared to mono antiplateletincluded 14 studies with 9012 patients, including the recent CHANCE 2012 study that was substantially larger than all previoustrials. Dual antiplatelet therapies included aspirin + clopidogrel, aspirin + dipyridamole and cilostazol + aspirin. Dual antiplatelettherapy produced significant reductions in the risk of recurrent stroke (RR 0.69, 95% CI 0.60 to 0.80) and the composite outcomeof stroke, TIA, ACS, and all deaths, with no significant increase in major bleeding. Sensitivity analyses that included only the 7 trialsthat were double-blinded showed similar results. While there was no significant heterogeneity across different dual and monoantiplatelet treatment regimens, the majority of data came from the aspirin + clopidogrel versus aspirin alone comparisoninvestigated in the CHANCE trial.

Two subsequent randomised controlled trials (He et al 2015 [41]; Yi et al 2014 [42]) compared combined clopidogrel and aspirintreatment to aspirin alone. Both had short follow-up periods, of 14 or 30 days. In both trials, neurological deterioration andrecurrent stroke occurred less often in the combined therapy groups, although only one of the studies reported that thesedifferences were statistically significant.

OutcomeTimeframe



Mono antiplatelettherapy

Dual antiplatelettherapy

Certainty ineffect


Summary

Recurrentstroke

7 days to 18+months of follow-

up

8 Critical

Relative risk 0.69(CI 95% 0.6 - 0.8)


(Randomized controlled)Follow up 7 days to 18+

months

90per 1000

62per 1000



indirectness

Dual antiplatelet therapyprobably decreases

recurrent stroke

Compositeoutcome of

stroke, TIA, ACS,and all deaths

7 days to 18+months of follow-

up

8 Critical

Relative risk 0.71(CI 95% 0.63 - 0.81)


(Randomized controlled)Follow up 7 days-18

months

121per 1000

86per 1000



indirectness

Dual antiplatelet therapyprobably decreases

composite outcome ofstroke, TIA, ACS, and all

deaths


75 of 157

References

[48] He F, Xia C, Zhang J-H, Li X-Q, Zhou Z-H, Li F-P, Li W, Lv Y, Chen H-S Clopidogrel plus aspirin versus aspirin alone forpreventing early neurological deterioration in patients with acute ischemic stroke.. Journal of clinical neuroscience : officialjournal of the Neurosurgical Society of Australasia 2015;22(1):83-6- Pubmed Journal

[49] Yi X, Lin J, Wang C, Zhang B, Chi W A comparative study of dual versus monoantiplatelet therapy in patients with acute large-artery atherosclerosis stroke.. Journal of stroke and cerebrovascular diseases : the official journal of National Stroke Association2014;23(7):1975-81- Pubmed Journal

[52] Wong KSL, Wang Y, Leng X, Mao C, Tang J, Bath PMW, Markus HS, Gorelick PB, Liu L, Lin W, Wang Y Early dual versus monoantiplatelet therapy for acute non-cardioembolic ischemic stroke or transient ischemic attack: an updated systematic review andmeta-analysis.. Circulation 2013;128(15):1656-66- Pubmed Journal

Major bleeding7 days to 18+

months of follow-up

8 Critical

Relative risk 1.35(CI 95% 0.7 - 2.59)


(Randomized controlled)Follow up 7 days to 18+

months

4per 1000

5per 1000



indirectness

dual antiplatelet therapyprobably has little or no

difference on majorbleeding


Recurrent stroke


Risk of bias: No seriousInconsistency: No seriousIndirectness: Serious Differences between the population of interest and those studied - themajority was Asian population ;Imprecision: No seriousPublication bias: No serious

Compositeoutcome of stroke,

TIA, ACS, and alldeaths



Major bleeding




76 of 157


http://dx.doi.org/10.1016/j.jocn.2014.05.038


http://dx.doi.org/10.1016/j.jstrokecerebrovasdis.2014.01.022


http://dx.doi.org/10.1161/CIRCULATIONAHA.113.003187

6 - Acute blood pressure lowering therapy

2017 guideline:


While there is strong evidence for lowering BP for secondary prevention (see 5.3 Blood pressure lowering), treatment within the first 48

hours remains controversial with both high and low BP found to negatively affect patient outcomes.248, 249 Large drops in BP appear to lead

to poorer outcomes while modest lowering interventions may produce benefits.250 BP should be lowered to <185 mmHg systolic bloodpressure (SBP) and <110 mmHg diastolic blood pressure (DBP) for ischaemic stroke patients eligible for thrombolysis (see 4.1 Thrombolysis).

There have been a number of studies investigating different agents for controlling BP. One Cochrane review (65 RCTs) of BP lowering within24 hours of an acute CVD event (six RCTs specific to stroke, all assessing calcium channel blockers, found no difference in mortality at 10 days

(RR 0.81, 95% CI 0.54–1.21).251 Another updated Cochrane review (12 RCTs) of BP lowering within one week of stroke found angiotensin-converting enzyme (ACE) inhibitors, angiotensin receptor antagonists, calcium channel blockers, clonidine and glyceryl trinitrate each

lowered BP while phenylephrine appeared to increase BP.252 There is no evidence that therapy reduces mortality or improves functional

outcomes.252

One RCT (25 sICH patients and 147 ischaemic stroke patients with SBP>160 mmHg) found that labetalol or lisinopril can effectively reduce

BP without adverse events. IV labetalol worked more rapidly than oral labetol and may have a better safety profile.253 No difference in deathor dependency at two weeks was found but 3-month mortality was halved (9·7% vs 20·3%, HR 0·40, 95% CI 0·2–1·0). The study wasunderpowered for clear outcomes data and a larger study is needed. Potential adverse events for those with ICH were also reported at two

weeks but numbers are too low (14 vs 3) for meaningful analysis.253 In the absence of clear data there was consensus that for patients withsevere hypertension, commencing or increasing BP therapy should be considered. Close monitoring of BP is recommended for all patients(see 4.7 Physiological monitoring).

There is currently insufficient evidence to recommend precise BP thresholds or targets in acute primary ICH. There is a general consensusthat severe sustained elevated BP (e.g. SBP> 180 mmHg) can be treated, especially if there is evidence or suspicion of raised intracranial

pressure. Evidence from the pilot phase of the INTERACT trial 254 has shown that more intensive BP lowering did not appear to be associatedwith any major hazard (e.g. haematoma growth) in the first few hours after stroke onset. The results of the INTERACT-2 and other trials willbe available in the coming years to determine clinical efficacy and help define the ideal BP targets for patients in the acute phase.

Practice Statement

• All acute stroke patients should have their blood pressure closely monitored in the first 48 hours after stroke onset.

• Patients with acute ischaemic stroke eligible for treatment with intravenous thrombolysis should have their blood pressure

reduced to below 185/110 mmHg before treatment and in the first 24 hours after treatment.

• Patients with acute ischaemic stroke with blood preesure >220/120/mmHg should have their blood pressure cautiously reduced

(e.g. by no more than 20%) over the first 24 hours.

Practical Info

Patients with ischemic stroke who are presented within 4.5 hours of onset should have their blood pressure managed aggressively,targeting a SBP < 180 mmHg anticipating potential intravenous thrombolytic therapy.

In patients with stroke, there is no urgency in continuing with anti-hypertensive medications unless the patient is medically stable and asafe route of administration established (either oral or NG).

Key Info

Benefits and harms Small net benefit, or little difference between alternatives


77 of 157

Rationale

Available evidence suggests high blood pressure in acute stroke is associated with poor outcome. Studies in blood pressure loweringtherapy in acute stroke however, have failed to show a benefit. Results from ongoing studies targeting the hyper-acute phase may answerthis important clinical question. Blood pressure lowering therapy, except for patients being considered for intravenous thrombolysis andin the case of extreme hypertension, cannot be recommended.

Available evidence from 8 studies with 11015 participants suggested acute blood pressure lowering therapy does not reduce death ordependency either by drug class, stroke subtype or time to treatment.

Quality of evidence

The quality of the studies included in the Cochrane review, which the recommendation on this topic is based on, was variable.

Moderate


None identified or expected


Resources and other considerations Factors not considered


Population: Adults with ischemic stroke

Intervention: Blood pressure lowering

Comparator: Control

Summary

Two systematic reviews from Lee et al (2015) [56] and Bath et al (2014) [57] showed that there was no overall effect of treatmenton death as an outcome in the studies analysed. No differences were observed when analysed by the subgroup of ischemic strokeeither .

OutcomeTimeframe



Control Blood pressurelowering

Certainty ineffect


Summary

Death anddependency

9 Critical

Odds Ratio 1(CI 95% 0.92 - 1.08)Based on data from11,015 patients in 8


409per 1000

409per 1000


HighBlood pressure loweringhas little or no differenceon death and dependency


Death anddependency

Intervention reference:Systematic review [57]

Risk of bias: No seriousInconsistency: No serious


78 of 157


Aggressive blood pressure lowering in the acute phase of care to a target SBP of <140mmHg is not recommended (Bath and Krishnan

2014 [57]; Qureshi et al. 2016 [58]; Lee et al. 2015 [56]).

Key Info

Rationale

High-quality evidence showed that there was no overall effect of acute BP lowering to <140mHg on death or functional outcome.

References

[56] Lee M, Ovbiagele B, Hong K-S, Wu Y-L, Lee J-E, Rao NM, Feng W, Saver JL Effect of Blood Pressure Lowering in Early IschemicStroke: Meta-Analysis . Stroke 2015/05/28 ; Website

[57] Bath PM, Krishnan K Interventions for deliberately altering blood pressure in acute stroke. Cochrane Database of SystematicReviews 2014; Journal

with included studies:CATIS 2013, CHHIPS2009, ENOS 2014, Eveson2007, PRoFESS 2009,RIGHT 2013, SCAST 2011,VENTURE 2013,Baseline/comparatorreference: Control arm ofreference used forintervention

Indirectness: No seriousImprecision: No seriousPublication bias: No serious

Benefits and harms

No benefits were found in a robust Cochrane systematic review of acute blood pressure lowering to SBP<140mmHg


Quality of evidence

The evidence has multiple high quality RCTs.

High




Resources and other considerations Important issues, or potential issues not investigated


79 of 157

http://stroke.ahajournals.org/content/early/2015/05/28/STROKEAHA.115.009552.abstract



Population: Adults with ICH


Comparator: Control

Summary

Evidence is the systematic review by Bath et al (2014) [57] which primarily comes from two large, well-designed RCTs examiningthe effect of acute blood pressure lowering in ICH over the last 5 years. One of them is INTERACT2, which suggest a systolictarget of 140mmHg probably improves outcomes, while another recent trial ATACH-II published after this systematic review didnot support aggressively lowering the SBP to less than 140mmHg (Qureshi et al 2016) [58].

References


[58] Qureshi AI, Palesch YY, Barsan WG, Hanley DF, Hsu CY, Martin RL, Moy CS, Silbergleit R, Steiner T, Suarez JI, Toyoda K, WangY, Yamamoto H, Yoon B-W Intensive Blood-Pressure Lowering in Patients with Acute Cerebral Hemorrhage.. The New Englandjournal of medicine 2016; Pubmed

OutcomeTimeframe




Certainty ineffect


Summary

Death anddependency

9 Critical

Odds Ratio 1.01(CI 95% 0.84 - 1.21)



543per 1000

545per 1000


High

In patients with mild tomoderate size ICH, a

treatment target of SBP140 has little or no

difference on death anddependency.


Death anddependency

Intervention reference:Systematic review [57]with included studies:CHHIPS 2009, ENOS2014, INTERACT pilot2008, INTERACT-2 2013,Koch 2008, RIGHT 2013,SCAST 2011,Baseline/comparatorreference: Control arm ofreference used forintervention



80 of 157




Population: Adults with ischemic stroke


Comparator: Control

Summary

Two systematic reviews from Lee et al (2015) [56] and Bath et al (2014) [57] showed that there was no overall effect of treatmenton death as an outcome in the studies analysed. No differences were observed when analysed by the subgroup of ischemic strokeeither .

References

[56] Lee M, Ovbiagele B, Hong K-S, Wu Y-L, Lee J-E, Rao NM, Feng W, Saver JL Effect of Blood Pressure Lowering in Early IschemicStroke: Meta-Analysis . Stroke 2015/05/28 ; Website


OutcomeTimeframe




Certainty ineffect


Summary

Death anddependency

9 Critical

Odds Ratio 1(CI 95% 0.92 - 1.08)Based on data from11,015 patients in 8


409per 1000

409per 1000


HighBlood pressure loweringhas little or no differenceon death and dependency


Death anddependency

Intervention reference:Systematic review [57]with included studies:CATIS 2013, CHHIPS2009, ENOS 2014, Eveson2007, PRoFESS 2009,RIGHT 2013, SCAST 2011,VENTURE 2013,Baseline/comparatorreference: Control arm ofreference used forintervention



81 of 157



Weak Recommendation

In people with acute primary intracerebral haemorrhage blood pressure may be cautiously reduced with a target SBP of, but not below,

140mmHg (Bath and Krishnan 2014 [57]; Qureshi et al. 2016 [58]).

Key Info

Rationale

Data from meta-analysis, together with result from the recent ATACH-2, suggests that in patients with mild to moderate ICH, an SBPtarget of 140mmHg (but not lower), is probably safe and associated with better patient outcomes as demonstrated by a shift in mRS at 90days.

Benefits and harms

The evidence of this recommendation is based on the Cochrane review by Bath, incorporating results from a large RCT INTERACT2(N = 2794). In INTERACT2, The primary end point of death or major disability at 3 months between the intensive treatment group andthe control group just fall short of reaching statistical significance (OR 0.87, 95% CI 0.75-1.01). An ordinal analysis of modified Rankinscores indicated improved functional outcomes with intensive lowering of blood pressure (OR 0.87, 95%CI 0.77 - 1.00). Results fromATACH-II did not support aggressively lowering the SBP to less than 140mmHg - there was no difference in death or disability but ahigher rate of serious adverse events.


Quality of evidence

Multiple high quality RCT

High








Comparator: Control

Summary


OutcomeTimeframe




Certainty ineffect

estimatesSummary


82 of 157

Weak Recommendation

Pre-existing antihypertensives may be withheld until patients are neurologically stable and treatment can be given safely (Bath and

Krishnan 2014 [57]).

Key Info

References



(Quality ofevidence)

Death anddependency

9 Critical

Odds Ratio 1.01(CI 95% 0.84 - 1.21)



543per 1000

545per 1000


High





Death anddependency



Benefits and harms

In the primary analysis of the ENOS study, continuing pre-stroke anti-hypertensives did not affect the primary outcome but wasassociated with worse Barthel Index at 90 days. The exact reason for this is uncertain.



83 of 157



Rationale

Based on limited available evidence, there appears to be no urgency in resuming pre-stroke anti-hypertensive therapy in acute strokepatients. Doing so may be associated with worsening functional outcome and it is advisable to wait until a safe route of administration isestablished.

Quality of evidence

High quality RCT data mainly from one study

High


Not identified and no variation in preference and values expected.





Intervention: Continue pre-stroke antihypertensives

Comparator: Stop pre-stroke antihypertensives

Summary

Bath et al (2014) [57] conducted a systematic review of the effectiveness of altering blood pressure in acute stroke patients . In atotal of 2860 patients, they did not find a significant difference of death or dependency between patients who continued pre-stroke anti-hypertensive treatment and whose who stopped. However, better functional outcomes measured by Barthel Indexwere associated with discontinuation of antihypertensives.

OutcomeTimeframe



Stop pre-strokeantihypertensives

Continue pre-strokeantihypertensives

Certainty ineffect


Summary

Death ordependency

9 Critical

Odds Ratio 1.06(CI 95% 0.91 - 1.24)



567per 1000

581per 1000


High

continue pre-strokeantihypertensives may

have little or nodifference on death or

dependency


Death ordependency

Intervention reference:Systematic review [57]with included studies:COSSACS 2010, ENOS2014,Baseline/comparator



84 of 157

References


reference: Control arm ofreference used forintervention


85 of 157


7 - Surgery for ischaemic stroke and management of cerebral oedema


Cerebral oedema in the infarcted or peri-lesional brain tissue often leads to early deterioration and death.255 Hemicraniectomy for ischaemicstroke should be considered for large life-threatening, space-occupying brain oedema or middle cerebral artery (MCA) infarcts where

prognosis is poor, so called ‘malignant infarction’.256 A meta-analysis (three RCTs) found decompressive surgery combined with medical

therapy reduced mortality and improved functional outcomes compared to medical therapy alone.256 These benefits were seen only inselected patients who fulfilled clear inclusion criteria (e.g. 18–60 years old, who can undergo surgery within 48 hours of symptom onset and

with clinical deficits suggesting significant MCA involvement).256 Given the prognosis for patients with malignant infarction, an urgentreferral for a neurosurgical opinion is strongly recommended.

One Cochrane review failed to find any RCTs for the use of angioplasty and stenting for intra-cranial artery stenosis.257 Evidence from case-series studies with three or more cases demonstrated an overall peri-operative rate of stroke of 7.9%, of death 3.4%, and of stroke or death9.5%. Further data are required before clear conclusions can be made regarding this intervention.

One Cochrane review (seven RCTs) found corticosteroid treatment has no benefit and may in fact cause harm and is therefore not

recommended.258

Another Cochrane review (11 RCTs) found osmotherapy with glycerol reduced short-term (~1 week) mortality in those with ischaemic stroke(OR 0.65, 95% CI 0.44–0.97) but this reduction was no longer significant when all strokes were considered (OR 0.78, 95% CI 0.58–1.06) or at

the end of scheduled trial follow-up (OR 0.98, 95% CI 0.73–1.31).259 Two RCTs found a non-significant improvement in a good outcome at theend of follow-up (OR 0.73, 95% CI 0.37–1.42). Osmotherapy should only be considered in selected cases (e.g. while assessing whether to

proceed with decompressive surgery).259

According to a Cochrane review (three small RCTs), there is insufficient evidence on the effects of mannitol in acute stroke.260

Hyperventilation has not been rigorously evaluated in stroke but short-term benefits have been found in patients with traumatic brain

injury.261


Selected patients aged 60 years and under with malignant middle cerebral artery territory infarction should undergo urgent

neurosurgical assessment for consideration of decompressive hemicraniectomy. When undertaken h e m i c r a n i e c t o m y surgery

should i d e a l l y be performed within 48 hours of symptom onset. (Cruz-Flores et al 2012 [61])

Needs rationale and practical advice (if any)

Key Info

Benefits and harms

There is a clear benefit in terms of survival. For those that survived there was a non-significant trend toward moderate to severedisability.


Quality of evidence

The overall quality of evidence is moderate with a meta-analysis of 3 randomized controlled trials in which numbers are small andconfidence intervals are wide. An individual patient pooled analysis of patients treated within 48 hours also supports most of thefindings.

Moderate

Preference and values Substantial variability is expected or uncertain


86 of 157

Most patients treated with hemicraniectomy will survive with at least long term moderate disability due to the underlying stroke, andthis should be discussed prior to treatment.


Hofmeijer et al (2013) [63] found that surgical decompression for space-occupying hemispheric infarction was not cost-effective perQALY gained compared to best medical treatment (ICER €60000 over a lifetime).

Factors not considered


Population: Adults < 60 y.o. with malignant middle cerebral artery infarct

Intervention: Hemicraniectomy

Comparator: Medical treatment

Summary

A 2012 Cochrane review by Cruz-Flores et al [61] included 3 trials (total N = 134) assessing the effectiveness of decompressivesurgery following acute ischaemic stroke with cerebral oedema. All 3 trials were restricted to patients aged 60 years or younger.Meta-analysis showed significant decreases in the risk of death (OR 0.19, 95% CI 0.09 to 0.37) and the risk of death or severedisability (modified Rankin scale scores > 4) at 12 months (OR 0.26, 95% CI 0.13 to 0.51). However, there was no significantdifference in death or disability defined as modified Rankin scores > 3, suggesting that patients that do survive tend to have atleast moderate disability. All 3 trials included in the review were stopped early, meaning the effect sizes found in the meta-analysismay be overestimated.

A systematic review of trials investigating hemicraniectomy for middle cerebral artery infarction by Back et al (2015) [59] found 6studies meeting inclusion criteria. This included the 3 trials in the Cruz-Flores et al review as well as 3 more recent trials that didnot restrict inclusion to patients < 60 years of age. The meta-analysis found similar reductions in mortality but also providedfurther evidence that surviving patients experience at least moderately severe disability.

OutcomeTimeframe



Medical treatment Hemicraniectomy

Certainty ineffect


Summary

Death at end offollow-up

12 months offollow-up

9 Critical

Odds Ratio 0.19(CI 95% 0.09 - 0.37)



631per 1000

245per 1000



imprecision

Hemicraniectomyprobably decreases death

at end of follow-up

Death ordisability

defined as mRS >3 at end offollow-up


Odds Ratio 0.56(CI 95% 0.27 - 1.15)



754per 1000

632per 1000



imprecision

Hemicraniectomyprobably slightly

decreases death ordisability defined as mRS

>3 at end of follow-up


87 of 157

9 Critical

Death or severedisability

defined as mRS >4 at 12 months


9 Critical

Odds Ratio 0.26(CI 95% 0.13 - 0.51)



662per 1000

337per 1000



imprecision

Hemicraniectomyprobably decreases death

or severe disabilitydefined as mRS >4 at end

of follow-up

Severe disabilityamong survivorsdefined as mRS 4

to 5 at 12months

12 months

9 Critical

Odds Ratio 2.45(CI 95% 0.92 - 6.55)



333per 1000

550per 1000



imprecision

Hemicraniectomyprobably increases

severe disability amongsurvivors defined as mRS

4 to 5 at 12 months


Death at end offollow-up

Intervention reference:Systematic review [61]with included studies:DECIMAL 2007, DESTINY2007, HAMLET 2009,Baseline/comparatorreference: Control arm ofreference used forintervention

Risk of bias: No seriousInconsistency: No seriousIndirectness: No seriousImprecision: Serious Low number of patients ;Publication bias: No serious

Death or disabilitydefined as mRS > 3at end of follow-up



Death or severedisability defined as

mRS > 4 at 12months

Intervention reference:Systematic review [61]with included studies:DECIMAL 2007, DESTINY2007, HAMLET 2009,Baseline/comparator



88 of 157

Weak Recommendation

Decompressive hemicraniectomy may be considered in highly selected patients over the age of 60 years, after careful consideration of

the premorbid functional status , and c o n s i d e r a t i o n o f w h e t h e r i t i s i n t h e patient p r e f e r e n c e s ' s b e s t l o n g - t e r m i n t e

r e s t s. (Back et al 2015 [59]; Jüttler et al. (2014) [60]; Cruz-Flores et al 2012 [61])

Needs rationale and practical advice

Key Info

References

[61] Cruz-Flores S, Berge E, Whittle IR Surgical decompression for cerebral oedema in acute ischaemic stroke. Cochrane Databaseof Systematic Reviews 2012; Pubmed Journal

reference: Control arm ofreference used forintervention

Severe disabilityamong survivors

defined as mRS 4 to5 at 12 months



Benefits and harms

There is evidence from a single randomised controlled trial that hemicraniectomy in patients over the age of 60 with malignant middlecerebral artery territory infarction improves the odds of survival with moderately severe disability.


Quality of evidence

The evidence is based on a single randomised controlled trial (Destiny 2)

Moderate


Approximately 1 third of survivors after hemicraniectomy in the >60 age group had severe disability (mRS=5, i.e. nursing home levelof care).

Substantial variability is expected or uncertain



89 of 157



Hofmeijer et al (2013) [63] found that surgical decompression for space-occupying hemispheric infarction was not cost-effective perQALY gained compared to best medical treatment (ICER €60000 over a lifetime).


Population: Adults > 60 y.o. with malignant middle cerebral artery infarct

Intervention: Hemicraniectomy

Comparator: Medical treatment

Summary

Jüttler et al (2014) [60] conducted a randomised trial (N = 112) of hemicraniectomy for patients with malignant middle-cerebral-artery infarction, which was stopped early when significant benefits were seen at 6 month follow-up. The trial included patients61 years of age or older (up to 80) as benefits of hemicraniectomy have previously been demonstrated for people 60 years oryounger. The control group received conservative treatment in an intensive care unit. Hemicraniectomy significantly increasedthe proportion of patients surviving without severe disability at 6 months (OR 2.91, 95% CI 1.06 to 7.49) and 12 months, althoughall patients in both groups had at least moderate disability (modified Rankin scores >= 3). Substantially fewer patients in thehemicraniectomy group had died by 12-month follow up.

A systematic review of trials investigating hemicraniectomy for middle cerebral artery infarction by Back et al (2015) [59] found 6studies meeting inclusion criteria. This included the trial by Jüttler et al. as well as earlier trials that had only included youngerpatients (18 to 60 years) and two other recent trials that also included older patients (up to 75 or 80 years). The meta-analysisshowed that hemicraniectomy significantly reduced the odds of death at 6 and 12 months, as well as the odds of having a modifiedRankin score of 2 at 12 months. The benefits of hemicraniectomy that had been reported in earlier reviews including patients 60years or younger appear to apply equally to patients older than 60 years, although only the trial by Jüttler et al. provides directevidence for this specific subgroup.

Note: Jüttler et al (2014) [60] did not report relative effect estimates for the outcomes at 12 months. The estimates reported herewere manually calculated based on the reported raw numbers of events.

OutcomeTimeframe



Medical treatment Hemicraniectomy

Certainty ineffect


Summary

Survival withoutsevere disability

at 6 months6 months

9 Critical

Odds Ratio 2.91(CI 95% 1.06 - 7.49)



159per 1000

355per 1000


ModerateDue to seriousimprecision -single study


survival without severedisability at 6 months

Survival at 12months

12 months

9 Critical

Odds Ratio 4.23(CI 95% 1.86 - 9.6)



242per 1000

575per 1000




survival at 12 months


90 of 157

Survival withoutsevere disability

at 12 months12 months

9 Critical

Odds Ratio 3.23(CI 95% 1.32 - 7.91)



161per 1000

383per 1000




survival without severedisability at 12 months

NeurologicalOutcome at 12

Months12 months

9 Critical

Measured by: NIHSSScale: 0-42 Lower betterBased on data from: 112


42points (Median)

22points (Median)

Difference: 20 fewerCI 95%


significantaccording to a

Wilcoxon test. Dueto serious

imprecision -single study

Hemicraniectomyprobably improves

neurological outcome at12 months


Survival withoutsevere disability at

6 months


Risk of bias: No seriousInconsistency: No seriousIndirectness: No seriousImprecision: Serious Only data from one study, Low number of patients ;Publication bias: No serious

Survival at 12months


Risk of bias: No serious Incomplete data and/or large loss to follow up ;Inconsistency: No seriousIndirectness: No seriousImprecision: Serious Only data from one study ;Publication bias: No serious

Survival withoutsevere disability at

12 months



NeurologicalOutcome at 12

Months




91 of 157


Corticosteroids are not recommended for management of patients with brain oedema and raised intracranial pressure. (Sandercock et al

2011 [62])

Needs rationale and practical advice. Meeting doc has as weak rec against so check

Key Info

References

[59] Back L, Nagaraja V, Kapur A, Eslick GD Role of decompressive hemicraniectomy in extensive middle cerebral artery strokes: ameta-analysis of randomised trials.. Internal medicine journal 2015;45(7):711-7- Pubmed Journal

[60] Jüttler E, Unterberg A, Woitzik J, Bösel J, Amiri H, Sakowitz OW, Gondan M, Schiller P, Limprecht R, Luntz S, Schneider H,Pinzer T, Hobohm C, Meixensberger J, Hacke W Hemicraniectomy in older patients with extensive middle-cerebral-artery stroke..The New England journal of medicine 2014;370(12):1091-100- Pubmed Journal

[61] Cruz-Flores S, Berge E, Whittle IR Surgical decompression for cerebral oedema in acute ischaemic stroke. Cochrane Databaseof Systematic Reviews 2012; Pubmed Journal

Benefits and harms

There is no evidence that corticosteroids are of beneift from the treatment of brain oedema and raised intracranial pressure in stroke.


Quality of evidence

The evidence comes from a Cochrane meta-analysis of 8 randomised trials.

High


Population: Corticosteroids for acute ischaemic stroke

Intervention: Corticosteroids

Comparator: Placebo

Summary

Sandercock et al (2011) [62] conducted a Cochrane review of the effectiveness of corticosteroids in acute ischaemic stroke,including 8 randomised trials (N = 466). Trials were double-blinded with placebo controls but details on randomisation andallocation concealment were generally unclear. Meta-analysis showed no significant difference in the odds of death by 12 months(OR 0.87, 95% CI 0.57 to 1.34) or within one month (OR 0.97, 95% CI 0.63 to 1.47). Due to the small numbers of included trialsand patients, the review authors noted that there is insufficient evidence to rule out benefit from corticosteroid treatment but atpresent there is evidence to support the use of corticosteroids.


92 of 157


http://dx.doi.org/10.1111/imj.12724





OutcomeTimeframe



Placebo Corticosteroids

Certainty ineffect


Summary

All deathsEnd of follow-up

9 Critical

Odds Ratio 0.87(CI 95% 0.57 - 1.34)


(Randomized controlled)Follow up 2 weeks to 12

months

379per 1000

347per 1000


HighCorticosteroids have

little or no difference onall deaths

Deaths withinone month

1 month

9 Critical

Odds Ratio 0.97(CI 95% 0.63 - 1.47)



281per 1000

275per 1000


HighCorticosteroids have

little or no difference ondeaths within one month


All deaths

Intervention reference:Systematic review [62]with included studies:Bauer 1973, Gupta 1978,McQueen 1978, Mulley1978, Norris 1976, Norris1986, Ogun 2001, Patten1972,Baseline/comparatorreference: Control arm ofreference used forintervention


Deaths within onemonth

Intervention reference:Systematic review [62]with included studies:Bauer 1973, Gupta 1978,McQueen 1978, Mulley1978, Norris 1976, Norris1986, Ogun 2001, Patten1972,Baseline/comparatorreference: Control arm ofreference used forintervention



93 of 157

Practice Statement

Osmotherapy and hyperventilation can be trialled while a neurosurgical consultation is undertaken.

References

[62] Sandercock PA, Soane T Corticosteroids for acute ischaemic stroke. Cochrane Database of Systematic Reviews 2011;Pubmed Journal


94 of 157



8 - Intracerebral haemorrhage (ICH) management

In general the management of ICH is similar to that for ischaemic stroke (e.g. rapid assessment, routine investigations, and prevention ofcomplications). This section addresses medical and surgical management specific to patients with ICH.

8.1 - Medical interventions

Haematoma growth is predictive of mortality and poor outcomes after ICH.262 A Cochrane review (six RCTs) found recombinant activated

factor VII (rFVIIa) reduced haematoma growth but did not reduce death or dependency at three months (RR 0.91, 95% CI 0.72–1.15).263 Theuse of rFVIIa in the treatment of ICH should be considered experimental and further trials are needed before recommendations on its

usefulness in routine clinical practice can be made.264

The neuroprotective agents that have been tested (e.g. gavestinel) have shown no clear benefits for patients with ICH.265 Citicoline has been

evaluated in a very small Phase I study and further larger studies are needed.266 Corticosteroids, glycerol and mannitol have all failed to show

benefits for patients with ICH.259, 260, 267

While there is consensus that ICH due to anticoagulation therapy should be urgently reversed there is no clear consensus about which

strategies to choose due to the lack of good quality data.268, 269 Traditional approaches include administration of fresh-frozen plasma (FFP)

and vitamin K, with prothrombin complex concentrate becoming more widely used in recent times.259, 268, 269 Off-label use of rFVIIa alone or

in combination with FFP has also been reported in small observational studies but is viewed as experimental only.270, 271

Management of BP is particularly important in ICH (see 4.4 Acute blood pressure lowering therapy) and is currently the subject of a largeAustralian RCT (INTERACT-2).

Weak Recommendation

1. Prothrombin complex concentrate should be used in preference to standard fresh frozen plasma to reverse coagulopathy in patients

with warfarin-related intracerebral haemorrhage.

2. Vitamin K should be used in addition to prothrombin complex to reverse warfarin but is insufficient as a sole treatment.

Please complete rationale and any practical advice

Key Info

Benefits and harms

Although prothrombin complex concentrate has clear superiority in rapid normalisation of coagluopathy, and probably reduces therisk of haematoma expansion (which is the rationale for treating), the effects on mortality and functional independence are less clear.


Quality of evidence

One well-conducted trial. The evidence for coagulopathy reversal is very robust, but evidence for haematoma expansion andmortality reduction is weaker due to sample size.

Moderate

Preference and values No substantial variability expected


PCC is more expensive than FFP and cost-utility is uncertain.

Important issues, or potential issues not investigated


95 of 157


Population: Adults with intracranial haemorrhage related to vitamin K antagonists

Intervention: Prothrombin complex concentrate

Comparator: Fresh frozen plasma

Summary

Steiner et al (2016) [61] conducted a randomised open label trial comparing fresh frozen plasma (FFP) to prothrombin complexconcentrate (PCC) for patients with intracranial haemorrhage related to vitamin K antagonists. The trial was terminated after 50patients had been included due to safety concerns, with greater haematoma expansion in the FFP group. Patients receiving PCCwere significantly more likely to have a normalised international normalised ratio (INR) within 3 hours (OR 30.6, 95% CI 4.7 to197.9) and showed significantly lower haematoma expansion. There were no significant differences in functional independence ordeath by 90 days. The early stopping of the trial suggests a risk of bias, particularly a risk that the differences in haematomaexpansion could be overestimated. The early stopping may also have limited the power of the study to detect differences inclinical outcomes.

Note: Steiner et al (2016) [61] did not report a relative effect estimate for deaths, instead reporting the results of a log-rank testbased on time-to-event data. The log-rank test was non-significant, p = 0.14. The relative risk reported here was manuallycalculated from the raw numbers of events.

OutcomeTimeframe



Fresh frozen plasma Prothrombincomplex concentrate

Certainty ineffect


Summary

INR ≤1·2 within3 h

3 hours

7 Critical

Odds Ratio 30.6(CI 95% 4.7 - 197.9)



87per 1000

745per 1000


LowDue to serious risk

of bias, Due toserious

imprecision

Prothrombin complexconcentrate may improve

the chances of INRreduction to ≤1·2 within

3 h in patients withwarfarin related ICH

Death90 days

9 Critical

Relative risk 0.53(CI 95% 0.2 - 1.4)



348per 1000

184per 1000




imprecision

prothrombin complexconcentrate maydecrease death in

patients with warfarinrelated ICH, compared

with standard FFP


90 days

8 Critical

Odds Ratio 1.7(CI 95% 0.4 - 6.8)



391per 1000

522per 1000




imprecision

prothrombin complexconcentrate may increasefunctional independence


96 of 157

Haematomaexpansion

24 hours

7 Critical

Measured by: blood in brain(mL)

Lower betterBased on data from: 50


Follow up 24 hours

mL (n/a) mL (n/a)

Difference: MD 16.4 fewer( CI 95% 2.9 fewer - 29.9 fewer )



imprecision

Prothrombin complexconcentrate may

decrease haematomaexpansion


INR ≤1·2 within 3 h


Risk of bias: Serious Missing intention-to-treat analysis, Trials stopping earlier thanscheduled, resulting in potential for overestimating benefits, Inadequate/lack of blinding ofparticipants and personnel, resulting in potential for performance bias ;Inconsistency: No seriousIndirectness: No seriousImprecision: Serious Wide confidence intervals, Only data from one study ;Publication bias: No serious

Death

Intervention reference:Primary study [68] freshfrozen plasma vers,Baseline/comparatorreference: Control arm ofreference used forintervention

Risk of bias: Serious Inadequate/lack of blinding of participants and personnel, resulting inpotential for performance bias, Trials stopping earlier than scheduled, resulting in potentialfor overestimating benefits ;Inconsistency: No seriousIndirectness: No seriousImprecision: Serious Wide confidence intervals, Only data from one study, Low number ofpatients ;Publication bias: No serious



Risk of bias: Serious Inadequate/lack of blinding of participants and personnel, resulting inpotential for performance bias, Inadequate concealment of allocation during randomizationprocess, resulting in potential for selection bias, Missing intention-to-treat analysis, Trialsstopping earlier than scheduled, resulting in potential for overestimating benefits ;Inconsistency: No seriousIndirectness: No seriousImprecision: Serious Low number of patients, Only data from one study, Wide confidenceintervals ;Publication bias: No serious

Haematomaexpansion


Risk of bias: Serious Inadequate/lack of blinding of participants and personnel, resulting inpotential for performance bias, Missing intention-to-treat analysis, Trials stopping earlierthan scheduled, resulting in potential for overestimating benefits ;Inconsistency: No seriousIndirectness: No seriousImprecision: Serious Low number of patients, Only data from one study, Wide confidenceintervals ;Publication bias: No serious


97 of 157


Patients with intracerebral hemorrhage related to DOACs should receive a specific reversal agent where available.

Add GRADE, rational, practical adviceLiterature search didn't identify direct evidence for this recommendations. Please inform Cindy if good evidence is available otherwise suggestchanging it to weak recommendation or practice statement.

Practical Info

(practical – idaruzicumab, future agents, promthrombinex imited data for Xa)

References

[68] Steiner T, Poli S, Griebe M, Hüsing J, Hajda J, Freiberger A, Bendszus M, Bösel J, Christensen H, Dohmen C, Hennerici M,Kollmer J, Stetefeld H, Wartenberg KE, Weimar C, Hacke W, Veltkamp R Fresh frozen plasma versus prothrombin complexconcentrate in patients with intracranial haemorrhage related to vitamin K antagonists (INCH): a randomised trial. The LancetNeurology 15(6):566-573 Journal Website


Population: Adults with intracranial haemorrhage related to vitamin K antagonists

Intervention: Prothrombin complex concentrate

Comparator: Fresh frozen plasma

Summary

Steiner et al (2016) [61] conducted a randomised open label trial comparing fresh frozen plasma (FFP) to prothrombin complexconcentrate (PCC) for patients with intracranial haemorrhage related to vitamin K antagonists. The trial was terminated after 50patients had been included due to safety concerns, with greater haematoma expansion in the FFP group. Patients receiving PCCwere significantly more likely to have a normalised international normalised ratio (INR) within 3 hours (OR 30.6, 95% CI 4.7 to197.9) and showed significantly lower haematoma expansion. There were no significant differences in functional independence ordeath by 90 days. The early stopping of the trial suggests a risk of bias, particularly a risk that the differences in haematomaexpansion could be overestimated. The early stopping may also have limited the power of the study to detect differences inclinical outcomes.

Note: Steiner et al (2016) [61] did not report a relative effect estimate for deaths, instead reporting the results of a log-rank testbased on time-to-event data. The log-rank test was non-significant, p = 0.14. The relative risk reported here was manuallycalculated from the raw numbers of events.


98 of 157

http://dx.doi.org/10.1016/S1474-4422(16)00110-1

http://dx.doi.org/10.1016/S1474-4422(16)00110-1

OutcomeTimeframe



Fresh frozen plasma Prothrombincomplex concentrate

Certainty ineffect


Summary

INR ≤1·2 within3 h

3 hours

7 Critical

Odds Ratio 30.6(CI 95% 4.7 - 197.9)



87per 1000

745per 1000




imprecision

Prothrombin complexconcentrate may improve

the chances of INRreduction to ≤1·2 within

3 h in patients withwarfarin related ICH

Death90 days

9 Critical

Relative risk 0.53(CI 95% 0.2 - 1.4)



348per 1000

184per 1000




imprecision

prothrombin complexconcentrate maydecrease death in

patients with warfarinrelated ICH, compared

with standard FFP


90 days

8 Critical

Odds Ratio 1.7(CI 95% 0.4 - 6.8)



391per 1000

522per 1000




imprecision

prothrombin complexconcentrate may increasefunctional independence

Haematomaexpansion

24 hours

7 Critical

Measured by: blood in brain(mL)



Follow up 24 hours

mL (n/a) mL (n/a)

Difference: MD 16.4 fewer( CI 95% 2.9 fewer - 29.9 fewer )



imprecision

Prothrombin complexconcentrate may

decrease haematomaexpansion


INR ≤1·2 within 3 h


Risk of bias: Serious Missing intention-to-treat analysis, Trials stopping earlier thanscheduled, resulting in potential for overestimating benefits, Inadequate/lack of blinding ofparticipants and personnel, resulting in potential for performance bias ;Inconsistency: No seriousIndirectness: No seriousImprecision: Serious Wide confidence intervals, Only data from one study ;Publication bias: No serious

Death

Intervention reference:Primary study [68] freshfrozen plasma vers,Baseline/comparatorreference: Control arm of

Risk of bias: Serious Inadequate/lack of blinding of participants and personnel, resulting inpotential for performance bias, Trials stopping earlier than scheduled, resulting in potentialfor overestimating benefits ;Inconsistency: No seriousIndirectness: No serious


99 of 157


In patients with intracerebral haemorrhage previously receiving antiplatelet therapy, we recommend that platelet transfusion should be

avoided.

Please complete Key info, Rationale and any practical advice

Key Info

References

[68] Steiner T, Poli S, Griebe M, Hüsing J, Hajda J, Freiberger A, Bendszus M, Bösel J, Christensen H, Dohmen C, Hennerici M,Kollmer J, Stetefeld H, Wartenberg KE, Weimar C, Hacke W, Veltkamp R Fresh frozen plasma versus prothrombin complexconcentrate in patients with intracranial haemorrhage related to vitamin K antagonists (INCH): a randomised trial. The LancetNeurology 15(6):566-573 Journal Website


Imprecision: Serious Wide confidence intervals, Only data from one study, Low number ofpatients ;Publication bias: No serious



Risk of bias: Serious Inadequate/lack of blinding of participants and personnel, resulting inpotential for performance bias, Inadequate concealment of allocation during randomizationprocess, resulting in potential for selection bias, Missing intention-to-treat analysis, Trialsstopping earlier than scheduled, resulting in potential for overestimating benefits ;Inconsistency: No seriousIndirectness: No seriousImprecision: Serious Low number of patients, Only data from one study, Wide confidenceintervals ;Publication bias: No serious

Haematomaexpansion


Risk of bias: Serious Inadequate/lack of blinding of participants and personnel, resulting inpotential for performance bias, Missing intention-to-treat analysis, Trials stopping earlierthan scheduled, resulting in potential for overestimating benefits ;Inconsistency: No seriousIndirectness: No seriousImprecision: Serious Low number of patients, Only data from one study, Wide confidenceintervals ;Publication bias: No serious

Benefits and harms

Increased rates of death and disability with treatment in a well conducted RCT, with plausible mechanism of harm, with consistentevidence of harm with both dichotomised MRS and shift analysis.



100 of 157

http://dx.doi.org/10.1016/S1474-4422(16)00110-1

http://dx.doi.org/10.1016/S1474-4422(16)00110-1

Quality of evidence Moderate




Population: Adults with intracerebral haemorrhage taking antiplatelet before

Intervention: Platelet transfusion


Summary

Baharoglu et al (2016) [62] conducted a multicentre open-label randomised trial (N = 190) of platelet transfusion after acuteintracerebral haemorrhage in people taking antiplatelet therapy. The intervention group received platelet transfusion within 6hours of intracerebral haemorrhage while the control group received standard care. While the trial was open label, outcomeassessors were blind to treatment allocation and allocation concealment was clearly reported. The primary analysis showedsignificantly increased odds of a shift towards death or dependence at 3 months (modified Rankin scale scores) following platelettransfusion (adjusted common OR 2.05, 95% CI 1.18 to 3.56). Patients receiving platelet transfusion also had significantlyincreased odds of a poor outcome at 3 months (mRS score 4-6, OR 2.04, 95% CI 1.12 to 3.74), with a nonsignificant decrease insurvival and increase in serious adverse events. These findings suggest platelet transfusion should not be used following acuteintracerebral haemorrhage.

OutcomeTimeframe



Standard care Platelet transfusion

Certainty ineffect


Summary

Death ordependence

90 days

9 Critical

Odds Ratio 2.04(CI 95% 1.12 - 3.74)



559per 1000

721per 1000



imprecision

Platelet transfusionprobably increases deathor dependence following

intracerebralhaemorrhage in patients

previously takingantiplatelet therapy

Survival90 days

9 Critical

Odds Ratio 0.62(CI 95% 0.33 - 1.19)



774per 1000

680per 1000



imprecision

Platelet transfusion maydecrease survival

following intracerebralhaemorrhage in patients

previously takingantiplatelet therapy


101 of 157


Aggressive blood pressure lowering in intracerebral hemorrhage to a target SBP of <140mmHg is not recommended.


References

[69] Baharoglu MI, Cordonnier C, Salman RA-S, de Gans K, Koopman MM, Brand A, Majoie CB, Beenen LF, Marquering HA,Vermeulen M, Nederkoorn PJ, de Haan RJ, Roos YB Platelet transfusion versus standard care after acute stroke due tospontaneous cerebral haemorrhage associated with antiplatelet therapy (PATCH): a randomised, open-label, phase 3 trial.. Lancet(London, England) 2016; Pubmed Journal

Serious adverseevents90 days

7 Critical

Odds Ratio 1.74(CI 95% 0.96 - 3.17)



295per 1000

421per 1000



imprecision

Platelet transfusion mayincrease serious adverse

events followingintracerebral

haemorrhage in patientspreviously taking

antiplatelet therapy


Death ordependence

Intervention reference:Primary studyBaseline/comparatorreference: Primary study

Risk of bias: No serious Inadequate/lack of blinding of participants and personnel, resultingin potential for performance bias ;Inconsistency: No seriousIndirectness: No seriousImprecision: Serious Only data from one study, 190 participants, Wide confidence intervals ;Publication bias: No serious

Survival


Risk of bias: No serious Inadequate/lack of blinding of participants and personnel, resultingin potential for performance bias ;Inconsistency: No seriousIndirectness: No seriousImprecision: Serious Wide confidence intervals, Only data from one study ;Publication bias: No serious

Serious adverseevents


Risk of bias: No seriousInconsistency: No seriousIndirectness: No seriousImprecision: Serious Wide confidence intervals, Only data from one study ;Publication bias: No serious


102 of 157


http://dx.doi.org/10.1016/S0140-6736(16)30392-0

Key Info

Rationale

Overall, the results of INTERACT2 suggest that in patients with mild to moderate ICH, an SBP target of 140mmHg is likely safe andassociated with small benefit in patient outcome as demonstrated by a shift in mRS at 90 days. Lowering below 140mmHg was not shownto be beneficial and may be associated with increased adverse effects.

Benefits and harms

The evidence of this recommendation came from INTERACT2. The primary end point of death or major disability at 3 monthsbetween the intensive treatment group and the control group just fell short of reaching statistical significance (p<0.06). A benefit wasseen in the 90 mRS between the groups in shift analysis (p<0.04). The ATACH2 trial showed no benefit (and trend to harm) of moreintensive BP lowering below 140mmHg.


Quality of evidence

High quality RCT

High





Intensive blood pressure therapy acutely is resource intensive both in terms of administration and monitoring. This interventiontypically requires an Emergency Department or High Dependency Unit/Intensive Care environment.





Comparator: Control

Summary


OutcomeTimeframe




Certainty ineffect


Summary

Death anddependency

Odds Ratio 1.01(CI 95% 0.84 - 1.21)

Based on data from 4,209543 545 High


treatment target of SBP


103 of 157

Weak Recommendation

In people with acute primary intracerebral haemorrhage blood pressure may be cautiously reduced with a target SBP of , b u t n o t b e l o

w , 140mmHg.


Key Info

References



9 Critical


per 1000 per 1000


140 has little or nodifference on death and

dependency.


Death anddependency



Benefits and harms

The evidence of this recommendation came from INTERACT2. The primary end point of death or major disability at 3 monthsbetween the intensive treatment group and the control group just fell short of reaching statistical significance (p<0.06). A benefit wasseen in the 90 mRS between the groups in shift analysis (p<0.04). The ATACH2 trial showed no benefit (and trend to harm) of moreintensive BP lowering below 140mmHg.



104 of 157



Rationale

Overall, the results of INTERACT2 suggest that in patients with mild to moderate ICH, an SBP target of 140mmHg is likely safe andassociated with small benefit in patient outcome as demonstrated by a shift in mRS at 90 days. Lowering below 140mmHg was not shownto be beneficial and may be associated with increased adverse effects.

Quality of evidence

High quality RCT

High





Intensive blood pressure therapy acutely is resource intensive both in terms of administration and monitoring. This interventiontypically requires an Emergency Department or High Dependency Unit/Intensive Care environment.





Comparator: Control

Summary


OutcomeTimeframe




Certainty ineffect


Summary

Death anddependency

9 Critical

Odds Ratio 1.01(CI 95% 0.84 - 1.21)



543per 1000

545per 1000


High





Death anddependency

Intervention reference:Systematic review [57]

Risk of bias: No seriousInconsistency: No serious


105 of 157

8.2 - Surgical interventions

Addition of the STICH II results [70] to prior meta-analyses gives 15 RCTs of surgery for ICH. These suggested better chances for a favourableoutcome with surgery (n=3366, OR 0.74, 95% CI 0.64-0.86) though with significant variability in the types of surgery and patients recruited.However, for the specific indication of lobar haemorrhage with no intraventricular extension there was no benefit from surgery (n=923, OR0.78, 95% CI 0.59-1.02) [70](53). Similarly, for patients with basal ganglia or thalamic haematomas surgery was not associated with higherrates of a favourable outcome (OR 0.84, 95% CI 0.65-110) [73]. Nor did surgery improve favourable outcome rates for patients withintraventricular haemorrhage complicating parenchymal haemorrhage (OR 0.77, 95% CI 0.45,1.31) [73].An individual patient data subgroup meta-analysis did suggest that certain clinical features may indicate a better response to surgery inpatients with supratentorial ICH [73]. These include surgery within 8 hours of ictus (p=0.003), haematoma volume 20-50 ml (p=0.004), GCSbetween 9 and 12 (p=0.0009) or patient age between 50 and 69 years (p=0.01).Other areas with less data but common practice include cerebellar haematomas e.g. if 'large' (>3 cm or so), posterior fossa decompression forhaemorrhagic (and ischaemic) stroke in that location, and EVD placement for supratentorial intracerebral haemorrhage with hydrocephaluswhen the hydrocephalus is considered to be a major contributor to the poor clinical state.The risks and benefits of surgery need to be carefully considered and balanced against the options that medical therapy can provide.

Practice Statement

In patients with large (>3cm) cerebellar haemorrhage, decompressive surgery may be offered to selected patients. For other

infratentorial haemorrhages (<3cm cerebellar, brainstem) the value of surgical intervention is unclear.

S u r g i c a l e v a c u a t i o n m a y b e u n d e r t a k e n f o r c e r e b e l l a r h e m i s p h e r e h a e m a t o m a s > 3 c m d i a m e t e r i n s e l e c t

e d p a t i e n t s .

References



with included studies:CHHIPS 2009, ENOS2014, INTERACT pilot2008, INTERACT-2 2013,Koch 2008, RIGHT 2013,SCAST 2011,Baseline/comparatorreference: Control arm ofreference used forintervention



106 of 157



Practical Info

Although there is little evidence to inform guideline development, there is a general consensus that patients with small haematomas andgood clinical status can be managed conservatively, and similarly patients with large haematomas with poor clinical status (e.g. low GCS)should be managed conservatively. In the middle are patients with moderate sized haematomas, and intermediate or deteriorating clinicalstatus who are often offered surgery for haematoma evacuation.In particular if the poor acute clinical state is considered to be primarily due to hydrocephalus rather than the effect of the haematoma onthe cerebellum and other posterior fossa structures, patients are often offered surgery for posterior fossa decompression (+/- haematomaevacuation).

Key Info

Rationale

There is little evidence to inform the appropriateness of haematoma evacuation for cerebellar haematomas.


Routine surgical evacuation of supratentorial intracerebral haemorrhage (lobar, basal ganglia and/or thalamic locations) is not

recommended outside the context of research .

Patients with lobar haematomas may be considered for surgery.

Practical Info

In patients with acute neurological deterioration considered to predominantly be due to obstructive hydrocephalus as a complication ofthe haematoma (as opposed to the haemorrhagic stroke itself), neurosurgical placement of an external ventricular drain is often offeredand is commonly accepted as beneficial although RCT evidence to support this is lacking.

Key Info

Benefits and harms

There is little evidence to suggest that surgery provides benefit for posterior fossa (brainstem and cerebellar) haemorrhages althoughcertain subgroups of patients (e.g. obstructive hydrocephalus, deteriorating neurological status) may derive benefit from surgery.


Quality of evidence

There are no randomised controlled trial data to inform this clinical scenario. The available evidence consists of case series only.

Very Low


In clinical situations when the neurological status is deteriorating and the outcome is likely to be fatal, it would be expected that manypeople (or substitute decision makers) would prefer to accept an offer of surgical management compared with palliation.



This therapy requires neurosurgical operative resources and is therefore more resource-intensive than conservative therapy.


Benefits and harms

There is probably benefit from surgery for supratentorial (lobar, basal ganglia and/or thalamic) haematomas.



107 of 157

Rationale

Although a meta-analysis suggested net benefit from surgery for lobar haematomas, there are several drawbacks including concernsabout the quality of the evidence and the resource-intensive nature of the intervention. Similarly, studies of surgery for basal ganglia and/or thalamic haematomas have reported nonsignificant results compared with conservative management. This therapy should therefore becarefully considered in each situation.

Quality of evidence

The quality of the evidence is poor. Although one meta-analysis demonstrates a statistically significant benefit from surgery,individual studies in this meta-analysis had non-overlapping confidence intervals. Furthermore another meta-analysis found nostatistically significant difference in outcomes.

Low


Some variation due to differences in the cultural or personal preferences of patients or substitute decision makers may be expected,but not significantly more variation than for other surgical procedures.



This treatment option would require ready access to neurosurgical services which may be challenging in some areas of Australia.



Population: Patients with Basal ganglia/thalamic haematomas

Intervention: Surgery

Comparator: Conservative treatment

Summary

An individual patient data meta-analysis by Gregson et al (2012) [73] compared surgery and conservative treatment in patientswith basal ganglia or thalamic haematomas. Of the eight studies included in the meta-analysis, three were of similar size (>190patients each) and the rest were smaller (<30 patients each). Of the three larger studies, two had point estimates suggestingoverall harm with surgery, although the CIs for the OR crossed the null 1.0. It was the third study which had a point estimateshowing benefit with surgery, with a CI for the OR that did not cross 1.0, which drove the overall point estimate of effect towardsbenefit with surgery, with a CI for the overall OR that crossed 1.0. Overall, surgery probably reduces unfavourable outcomes.

OutcomeTimeframe



Conservativetreatment

Surgery

Certainty ineffect


Summary

Unfavourableoutcome

3-6 months

Odds Ratio 0.84(CI 95% 0.65 - 1.1)


782per 1000

751per 1000

Difference: 31 fewer per 1000


Surgery probably reducesthe likelihood of an

unfavourable outcome inpatients with basal


108 of 157

References

[73] Gregson BA, Broderick JP, Auer LM, Batjer H, Chen X-C, Juvela S, Morgenstern LB, Pantazis GC, Teernstra OPM, Wang W-Z,Zuccarello M, Mendelow AD Individual patient data subgroup meta-analysis of surgery for spontaneous supratentorialintracerebral hemorrhage.. Stroke; a journal of cerebral circulation 2012;43(6):1496-504- Pubmed Journal

9 Critical

(Randomized controlled)Follow up 3-6 months

( CI 95% 16 more - 82 fewer )ganglia and thalamic

haematomas.


Unfavourableoutcome


Risk of bias: No serious Unable to tell from meta-analysis (Gregson) ;Inconsistency: Serious Point estimates vary widely,The confidence interval of some of thestudies do not overlap with those of most included studies/ the point estimate of some of theincluded studies., The magnitude of statistical heterogeneity was high, with I^2:70.8 %., , Thedirection of the effect is not consistent between the included studies. The point estimate forthe overall result is in favour of the intervention group, but this appears to be driven by asingle study (Wang), with all of the other studies' point estimates suggesting outcomes lessfavourable then reported by Wang et al. with intervention. ;Indirectness: No seriousPublication bias: No serious Smaller studies did not have particularly favourable outcomeswith surgery therefore probably not biased against publication for neutral/negative studies. ;


Population: Patients with lobar haematoma



Summary

A meta-analysis by Mendelow et al (2013) [70] showed that in patients with lobar haematomas, surgery probably reduces the rateof an unfavourable outcome slightly compared with initial conservative treatment (OR 0.74, 95%CI 0.64-0.86). In the meta-analysis by Mendelow et al., the confidence intervals for several of the contributing studies did not overlap, reducing the degree ofprecision of the estimate of effect. Conversely, crossover between the immediate surgery and delayed surgery groups may reducethe apparent impact of surgery on this outcome.

In a randomised trial by Xiao et al (2012) [71], patients (N = 36) with large (>70ml) lobar haematomas had CT-based haematomapuncture and aspiration (removing, an average of 1/3 of the haematoma volume) prior to haematoma evacuation via craniectomy.Survival at 12 months was better in those who had prior puncture and aspiration (58.3%) compared with patients who only hadcraniectomy without prior puncture and aspiration (20.8%). However, patients in the puncture and aspiration group had theircraniectomy on average 60 mins earlier than the group that proceeded directly to craniectomy, introducing the possibility that theimproved survival was related to earlier surgery rather than initial haematoma puncture and aspiration.


109 of 157



Overall, although there is some degree of uncertainty, surgery may reduce unfavourable outcomes in patients with lobarhaematoma.

References

[70] Mendelow AD, Gregson BA, Rowan EN, Murray GD, Gholkar A, Mitchell PM Early surgery versus initial conservativetreatment in patients with spontaneous supratentorial lobar intracerebral haematomas (STICH II): a randomised trial.. Lancet(London, England) 2013;382(9890):397-408- Pubmed Journal

[71] Xiao BO, Wu F-F, Zhang H, Ma Y-B A randomized study of urgent computed tomography-based hematoma puncture andaspiration in the emergency department and subsequent evacuation using craniectomy versus craniectomy only.. Journal ofneurosurgery 2012;117(3):566-73- Pubmed Journal

OutcomeTimeframe




Surgery

Certainty ineffect


Summary

Unfavourableoutcome

Not specified

8 Critical

Odds Ratio 0.74(CI 95% 0.64 - 0.86)


(Randomized controlled)Follow up Unclear

665per 1000

595per 1000


LowDue to serious

inconsistency, Dueto serious

indirectness

Surgery for patients withlobar haematomas may

reduce the rate ofunfavourable outcome

slightly.


Unfavourableoutcome


Risk of bias: No serious Difficult to determine from the meta-analysis (Mendelow et al.) ;Inconsistency: Serious The confidence interval of some of the studies do not overlap withthose of most included studies/ the point estimate of some of the included studies., Thedirection of the effect is not consistent between the included studies, The magnitude ofstatistical heterogeneity was high, with I^2:67 %. ;Indirectness: Serious Differences between the intervention/comparator of interest andthose studied (due to a variety of different interventions included in the meta-analysis) ;Imprecision: No seriousPublication bias: No serious


110 of 157


http://dx.doi.org/10.1016/S0140-6736(13)60986-1


http://dx.doi.org/10.3171/2012.5.JNS111611


Use of intraventricular thrombolysis via external ventricular drain catheter for the treatment of intraventricular haemorrhage is not

recommended outside the context of research.

Please review and edit Key info, Rationale, and any Practical advice

Key Info

Rationale

There are few studies assessing intraventricular thrombolysis or endoscopic surgery for ventricular haemorrhage. These studies includesmall numbers of patients, have variable outcome measures and do not demonstrate long term clinical benefit from such interventions.

Benefits and harms

Intraventricular haemorrhage thrombolysis is not recommended. Evidence does not demonstrate improved clinical outcomes, andsuggests increased risk of symptomatic haemorrhage.

Similarly, endoscopic surgery for intraventricular haemorrhage is not recommended. Evidence does not demonstrate improvedclinical outcomes.


Quality of evidence Low

Preference and values Substantial variability is expected or uncertain

Resources and other considerations Important issues, or potential issues not investigated


Population: Adults with intraventricular haemorrhage complicating parenchymal haemorrhage



Summary

In patients with intraventricular haemorrhage complicating parenchymal haemorrhage, a meta-analysis by Gregson et al (2012)[73] showed that surgery probably reduces the rate of unfavourable outcome. A small RCT (N = 48) by Chen et al (2010)[75] investigated endoscopic surgery compared with external ventricular drainage surgery for intraventricular haemorrhagecaused by thalamic haemorrhage. However, it showed little difference in critical clinical outcomes such as death and disability.

OutcomeTimeframe


Absolute effect estimatesCertainty in

effectestimates

Summary


111 of 157

References

[73] Gregson BA, Broderick JP, Auer LM, Batjer H, Chen X-C, Juvela S, Morgenstern LB, Pantazis GC, Teernstra OPM, Wang W-Z,Zuccarello M, Mendelow AD Individual patient data subgroup meta-analysis of surgery for spontaneous supratentorialintracerebral hemorrhage.. Stroke; a journal of cerebral circulation 2012;43(6):1496-504- Pubmed Journal

[75] Chen C-C, Liu C-L, Tung Y-N, Lee H-C, Chuang H-C, Lin S-Z, Cho D-Y Endoscopic surgery for intraventricular hemorrhage(IVH) caused by thalamic hemorrhage: comparisons of endoscopic surgery and external ventricular drainage (EVD) surgery..World neurosurgery 2011;75(2):264-8- Pubmed Journal


Surgery (Quality ofevidence)

Unfavourableoutcome

Unclear

8 Critical

Relative risk 0.77(CI 95% 0.45 - 1.31)


(Randomized controlled)Follow up Various

888per 1000

684per 1000



imprecision(statistically

nonsignificantoutcome)

Surgery forintraventricular

haematomascomplicatingparenchymal

haematomas possiblydecreases the rate of

unfavourable outcome.


Unfavourableoutcome


Risk of bias: No serious Difficult to tell from meta-analysis ;Inconsistency: No serious I^2=0 ;Indirectness: No seriousImprecision: Serious Wide confidence intervals ;Publication bias: No serious


Population: Adults with intraventricular haemorrhage complicating parenchymal haemorrhage

Intervention: Intraventricular thrombolysis

Comparator: Placebo

Summary

Naff et al (2011) [74] investigated the use of rtPA for intracerebral haemorrhage in a randomised controlled trial (N = 48) andshowed potential small benefits and large adverse effects. A small RCT (N = 16) by King et al (2012) [72] used intraventricularurokinase but no statistically significant differences were found in 6-month mortality, 30-day NIHSS score and 30-day mRS score.


112 of 157




http://dx.doi.org/10.1016/j.wneu.2010.07.041

OutcomeTimeframe



Placebo Intraventricularthrombolysis

Certainty ineffect


Summary

Death30 days

8 Critical

Relative risk 0.85(CI 95% 0.28 - 2.55)



227per 1000

193per 1000



imprecision

Intraventricular rtPAadministration probablyhas little or no effect on

mortality in patients withlarge ventricular

haemorrhages due toextension of spontaneous

small supratentorialintracranial

haemorrhage.

Adverse events -Ventriculitis

30 days

7 Critical

Relative risk 0.85(CI 95% 0.12 - 5.52)



91per 1000

77per 1000



imprecision

intraventricularthrombolysis probably

has little or no differenceon adverse events -

ventriculitis

Adverse events -Symptomatic

bleeding30 days

7 Critical

Relative risk 5.08(CI 95% 0.66 - 39.02)



45per 1000

229per 1000

Difference: 184 more per 1000( CI 95% 15 fewer - 1,711 more )


imprecision

intraventricularthrombolysis probably

increases adverse events- symptomatic bleeding


Death


Risk of bias: No seriousInconsistency: No seriousIndirectness: No seriousImprecision: Serious Low number of patients, Only data from one study, Wide confidenceintervals ;Publication bias: No serious

Adverse events -Ventriculitis


Risk of bias: No seriousInconsistency: No seriousIndirectness: No seriousImprecision: Serious Wide confidence intervals, Low number of patients, Only data from onestudy ;Publication bias: No serious


113 of 157

References

[72] King NKK, Lai JL, Tan LB, Lee KK, Pang BC, Ng I, Wang E A randomized, placebo-controlled pilot study of patients withspontaneous intraventricular haemorrhage treated with intraventricular thrombolysis.. Journal of clinical neuroscience : officialjournal of the Neurosurgical Society of Australasia 2012;19(7):961-4- Pubmed Journal

[74] Naff N, Williams MA, Keyl PM, Tuhrim S, Bullock MR, Mayer SA, Coplin W, Narayan R, Haines S, Cruz-Flores S, Zuccarello M,Brock D, Awad I, Ziai WC, Marmarou A, Rhoney D, McBee N, Lane K, Hanley DF Low-dose recombinant tissue-type plasminogenactivator enhances clot resolution in brain hemorrhage: the intraventricular hemorrhage thrombolysis trial.. Stroke; a journal ofcerebral circulation 2011;42(11):3009-16- Pubmed Journal

Adverse events -Symptomatic

bleeding


Risk of bias: No seriousInconsistency: No seriousIndirectness: No seriousImprecision: Serious Wide confidence intervals, Low number of patients, Only data from onestudy ;Publication bias: No serious


114 of 157





9 - Oxygen therapy


An updated Cochrane review (six RCTs) of hyperbaric oxygen therapy concluded that there are no significant differences in the death rate at

six months (RR 0.61, 95% CI 0.17–2.2, p=0.45).281 One quasi-RCT found no benefits of routine low-oxygen therapy for the first 24 hours in

stroke patients.282 A small RCT of eight hours of high-flow normobaric oxygen therapy started within 12 hours of onset in patients withperfusion-diffusion ‘mismatch’ on MRI found short-term improvements in stroke severity scales but no difference in patient outcomes at

three months.283 A recent RCT found that low-level nocturnal normobaric oxygen therapy commencing within 72 hours of stroke onsetincreased mean nocturnal oxygen saturation by 2.5% and reduced episodes of desaturation but produced no difference in any other

outcomes.284

Many centres represented in the stroke unit trials data had management policies for oxygen therapy.41 It was the consensus of the EWG thatpatients found to be hypoxic (<95% oxygen saturation) at any time (i.e. from pre-hospital to post-acute) should be given supplemental oxygen.


Patients who are hypoxic (i.e. <95% oxygen saturation) should be given supplemental oxygen.

PT: there is no direct evidence for this and it was a practice statement in 2010 guideline - suggest changing this to practice statement and completethe Rationale as needed.

Key Info

Rationale

There is inadequate evidence of benefit of supplemental oxygen therapy in normoxic ischaemic stroke patients. Many other non-strokestudies have found benefits in supplemental oxygen therapy for those who are hypoxic and it was the strong opinion of the working partythat such benefits are likely in those with stroke.


The routine use of supplemental oxygen is not recommended in acute stroke patients who are not hypoxic.


Benefits and harms Small net benefit, or little difference between alternatives

Quality of evidence Low



Supplemental oxygen equipment is costly but usually available in hospital .



115 of 157

Key Info

Rationale

Low quality evidence shows ambivalent results for routine use of oxygen supplementation for acute stroke patients. Considering the extracost inccured and uncertainty in the benefit, routine use of supplemental oxygen cannot be recommended.

Benefits and harms

Routine oxygen supplementation was shown to improve neurological outcome measured on NIHSS but not critical outcomes of deathand disability.


Quality of evidence

The overall quality is low due to high risk of bias and imprecision (data were from one single RCT).

Low


Routine use of oxygen is likely to incur extra cost.




Intervention: Early routine oxygen supplementation

Comparator: Room air

Summary

A pilot RCT with 289 participants showed no benefits in death or disability, but improvement in neurological outcomes at oneweek (Ali et al 2014 [77]; Roffe et al 2011 [78]) . The full trial of this pilot study (N=8003) has only been published in abstractversion and concluded that routine oxygen therapy does not improve functional outcome at 90 days in any of the predefinedsubgroups including stroke type or severity. A previous quasi-RCT with 550 patients also did not find routine supplementaloxygen to be beneficial in nonhypoxic stroke patients (Ronning et al 1999 [79]). Overall, the current evidence does not support theuse of routine oxygen supplementation.

OutcomeTimeframe



Room air Early routine oxygensupplementation

Certainty ineffect


Summary

Death1 week

9 Critical

Odds Ratio 1.2(CI 95% 0.32 - 4.55)



28per 1000

33per 1000


LowDue to serious riskof bias and serious

imprecision

Early routine oxygensupplementation may

have little or nodifference on death

Death at 6months

6 months

Hazard Ratio 1.1(CI 95% 0.59 - 2.07)


LowDue to serious

imprecision andserious risk of bias

early routine oxygensupplementation may

have little or no


116 of 157

9 Critical


difference on death at 6months

Improvement inneurological

outcome1 week

7 Critical

Odds Ratio 2.9(CI 95% 1.59 - 5.4)


(Randomized controlled)Follow up 1 week

135per 1000

312per 1000


LowDue to serious

imprecision, Dueto serious risk of

bias

Early routine oxygensupplementation mayimprove neurological

outcome at 1 week

Disability (mRS>= 3)

6 months

8 Critical

Odds Ratio 1.01(CI 95% 0.62 - 1.65)



554per 1000

556per 1000


LowDue to serious


bias

early routine oxygensupplementation may

have little or nodifference on disability

(mrs >= 3)

QOL6 months

7 Critical

Based on data from 222patients in 1 studies

Adjusted mean scores favoured the oxygengroup but the difference was not statistically

significant

Very LowDue to very

serious risk of bias,Due to serious

imprecision

We are uncertainwhether early routine

oxygen supplementationimproves or worsen QOL


Death


Risk of bias: Serious Single blinded trial ;Inconsistency: No seriousIndirectness: No seriousImprecision: Serious Only data from one study ;Publication bias: No serious

Death at 6 months


Risk of bias: Serious Single blinded trial ;Inconsistency: No seriousIndirectness: No seriousImprecision: Serious Only data from one study, Wide confidence intervals ;Publication bias: No serious

Improvement inneurological

outcome

Intervention reference:Primary study [78]roffe_2011_the sos pilot,Baseline/comparatorreference: Control arm ofreference used forintervention

Risk of bias: Serious single blinded trial ;Inconsistency: No seriousIndirectness: No seriousImprecision: Serious Only data from one study ;Publication bias: No serious


117 of 157


Hyperbaric oxygen therapy in acute ischaemic stroke patients is not recommended.


Key Info

References

[77] Ali K, Warusevitane A, Lally F, Sim J, Sills S, Pountain S, Nevatte T, Allen M, Roffe C The stroke oxygen pilot study: arandomized controlled trial of the effects of routine oxygen supplementation early after acute stroke--effect on key outcomes atsix months.. PloS one 2014;8(6):e59274- Pubmed Journal

[78] Roffe C, Ali K, Warusevitane A, Sills S, Pountain S, Allen M, Hodsoll J, Lally F, Jones P, Crome P The SOS pilot study: a RCT ofroutine oxygen supplementation early after acute stroke--effect on recovery of neurological function at one week.. PloS one2011;6(5):e19113- Pubmed Journal

[79] Rønning OM, Guldvog B Should stroke victims routinely receive supplemental oxygen? A quasi-randomized controlled trial..Stroke; a journal of cerebral circulation 1999;30(10):2033-7-null Pubmed

Disability (mRS >=3)

Intervention reference:Primary study [77]ali_2014_the stroke oxyg,Baseline/comparatorreference: Control arm ofreference used forintervention

Risk of bias: Serious single blinded trial ;Inconsistency: No seriousIndirectness: No seriousImprecision: Serious Only data from one study, Wide confidence intervals ;Publication bias: No serious

QOLIntervention reference:Primary study

Risk of bias: Very Serious Incomplete data and/or large loss to follow up, Inadequate/lack ofblinding of participants and personnel, resulting in potential for performance bias ;Inconsistency: No seriousIndirectness: No seriousImprecision: Serious Only data from one study; heterogeneity did not allow pooled data ;Publication bias: No serious

Benefits and harms

No benefit of hyperbaric oxygen was found in the outcomes of death or functional outcome in acute ischaemic stroke patients.


Quality of evidence

The overall quality of evidence is low for the outcome of death (downgraded due to small sample size and serious risk of bias) but verylow for functional outcome (due to very serious risk of bias, and inconsistency in results and in measurement).

Low


118 of 157


http://dx.doi.org/10.1371/journal.pone.0059274




Rationale

Low quality evidence shows ambivalent results for hyperbaric oxygen therapy for acute stroke patients. Considering extra cost incurredand uncertainty in the benefit, routine use of supplemental oxygen cannot be recommended.


Hyperbaric oxygen is expensive and resource intensive.




Intervention: Hyperbaric oxygen therapy

Comparator: Standard practice

Summary

Bennet et al (2014) [76] conducted a Cochrane review of hyperbaric oxygen therapy in acute stroke patients. They did not findevidence of improved clinical outcomes. However the overall quality of evidence was insufficient to exclude the possibility ofclinical benefits and well-designed studies in the future can provide a clearer answer.

OutcomeTimeframe



Standard practice Hyperbaric oxygentherapy

Certainty ineffect


Summary

Death3-6 months

9 Critical

Relative risk 0.97(CI 95% 0.34 - 2.75)


(Randomized controlled)Follow up 3- to 6-months

85per 1000

82per 1000


LowDue to serious


bias

hyperbaric oxygentherapy may have little or

no difference on death

Functionaloutcome

1 to 365 days

7 Critical

Based on data from 705patients in 11 studies

Four of 14 scale measures of disability andfunctional performance indicated

improvement following HBOT.

Very LowDue to very

serious risk of biasand serious

inconsistency

We are uncertainwhether hyperbaric

oxygen therapy increasesor decreases functional

outcome


Death

Intervention reference:Systematic review [76]with included studies:Baseline/comparatorreference: Control arm ofreference used forintervention

Risk of bias: Serious Inadequate concealment of allocation during randomization process,resulting in potential for selection bias ;Inconsistency: No seriousIndirectness: No seriousImprecision: Serious Small sample sizes ;Publication bias: No serious


119 of 157

References

[76] Bennett MH, Weibel S, Wasiak J, Schnabel A, French C, Kranke P Hyperbaric oxygen therapy for acute ischaemic stroke.. TheCochrane database of systematic reviews 2014;11 CD004954- Pubmed Journal

Functionaloutcome

Intervention reference:Systematic review

Risk of bias: Very Serious These trials varied in methodological quality, and only six providedfull reports of completed trials in a peer-reviewed publication, Inadequate concealment ofallocation during randomization process, resulting in potential for selection bias, Inadequate/lack of blinding of participants and personnel, resulting in potential for performance bias,Inadequate/lack of blinding of outcome assessors, resulting in potential for detection bias,Incomplete data and/or large loss to follow up ;Inconsistency: Serious The direction of the effect is not consistent between the includedstudies, and various measurement tools and timeframes ;Indirectness: No seriousImprecision: No serious no data pooled for this outcome ;Publication bias: No serious


120 of 157



10 - Neuroprotection

2017 guideline:

A large number of neuroprotective agents have been studied in clinical trials; however, none have demonstrated clear benefits and hencecannot be recommended for routine use.298-301 he most recent agent studied in a large trial, NXY-059, failed to show any benefits.302,303,[80] [81][82][83][84]

There are too few data on other groups of agents including colony-stimulating factors (including erythropoietin, granulocyte colony-stimulating factor and analogues),304 theophylline, aminophylline, and caffeine and analogues305 and further trials are required before clearconclusions can be drawn. A number of initial small trials have found potential benefits for albumin,306 edaravone,307 minocycline308 andarundic acid (ONO2506)309 but larger trials are required to confirm these preliminary results. iticoline may improve the chance of a goodrecovery at three months (OR 1.38, 95% CI 1.10–1.72) 310; a further large Phase III trial is ongoing.Recent studies have assessed the feasibility of reducing body temperature (via physical cooling or paracetamolhysical coolingacetaminophen)as an acute intervention. While such interventions appear promising 303, 311-314, a Cochrane review (eight RCTs/CCTs) found cooling viapharmacological or physical or pharmacological methods does not reduce the combined risk of death or dependency (OR 0.9, 95% CI 0.6–1.4)or death alone (OR 0.9, 95% CI 0.5–1.5). Both methods were associated with a non-significant increase in the occurrence of infections.315A subsequent large RCT (n=1400)316 of paracetamol (6 g) <12 hours in all patients with temperature 36–39°C showed a non-significanttrend to improved outcomes (OR 1.20, 95% CI 0.96–1.50). Post hoc analysis in those with temperatures of 37–39°C found significantimprovement indicating paracetamol may be appropriate only where fever occurs rather than routinely applied to all stroke patients (see 4.11Pyrexia management). Further robust trials are needed, particularly for physical cooling.Observational studies suggest that receiving statin therapy prior to stroke may have a neuroprotective effect. One small RCT in patients withischaemic stroke (n=89) compared the effect of continuing statin therapy (20 mg/day atorvastin) to ceasing therapy in the acute phase (firstthree days). Statin withdrawal was found to be associated with a 4.7-fold increase in the combined risk of death or dependency at threemonths and an 8.67-fold increase in the risk of early neurological deterioration.317 In contrast, a study of commencing new statin therapy(simvastatin 40 mg/day for the first week, then 20 mg/day) for patients admitted within 3–12 hours found no difference in biological markersor function at 90 days.318 While there was a significant reduction in impairment (>4 NIHSS, 46.4% vs 17.9%) by the third day of treatment,there was also a non-significant increase in mortality and a greater incidence of infections (OR 2.4, 95% CI 1.06–5.4) in the simvastatingroup.319 Further large interventional studies are needed to clarify the role of continuing or commencing statin therapy in acute strokepatients.

Practice Statement

Putative neuroprotective agents, including hypothermic cooling, should only be used in a randomised controlled trial.

Practice Statement

Patients with acute ischaemic stroke who were receiving statins prior to admission can continue statin treatment.


121 of 157

11 - Glycaemic therapy

Hyperglycaemia after stroke is found in one-third of patients although the reported incidence varies between 8% and 83% depending on the

cohort and definition (Capes et al 2001 [90]). Previously undetected diabetes is found in 16–24% of patients admitted with stroke (Gray et al2004 [91]; Kernan et al 2005 [92]). Observational data indicate that hyperglycaemia fluctuates in the first 72 hours in both non-diabetic anddiabetic patients even with current best practice (Allport et al 2006 [94]). Observational data also reveal poorer outcomes for non-diabeticpatients with hyperglycaemia (Capes et al 2001 [90]). Glucose intolerance after stroke is also common (approximately 25%) Kernan et al2005 [92]; Allport et al 2006 [94]) and linked to higher stroke recurrence (see Diabetes management) (Vermeer et al 2006 [95]).

There is now good evidence that hyperglycaemia needs management regardless of the patients diabetic status (Bellolio et al 2014 [85]; Ntaioset al 2014 [86]; Middleton et al 2011 [89] and Drury et al 2014 [98]). Implementation of effective glycaemic control requires education ofnursing staff across all shifts, which can be challenging. Glucometers also need to be readily available.


All patients should have their blood glucose level monitored and appropriate glycaemic therapy instituted to treat hyperglycaemia(glucose levels greater than 10mmols/L) regardless of the patients diabetic status. (Bellolio et al 2014 [85]; Ntaios et al 2014 [86];Middleton et al 2011 [89] and Drury et al 2014 [98])

Practical Info

The trials in the Cochrane review (Bellolio et al 2014) used tight control of blood glucose (4 - 7.5mmol/L) however the QASC trial(Middleton et al 2011) suggested that insulin should only be used to maintain BGL of less than 11mmol/L (euglycaemia) as part of a carebundle.

The Australian Diabetes Society Guidelines for Routine Glucose Control in Hospital recommend:

1. Patients admitted to hospital with acute thrombotic stroke who have hyperglycaemia, should betreated to achieve and maintain glucose levels less than 10 mmol/L.2. Hypoglycaemia must be avoided, and therefore it would be prudent to avoid treatment which lowersthe glucose below 5 mmol/L.

Key Info

Benefits and harms

The risk of hypoglycaemia was higher in the intervention groups treated with IV insulin (maintain glycaemic level between 4 - 7.5mmol/L) when compared with SC insulin or normal saline. This applied to symptomatic and non-symptomatic hypoglycaemia.

IV insulin has little or no difference on mortality or functional outcome when compared to SC insulin or normal saline.

IV insulin has little or no difference on dependency, death, indepence in daily activities or NIHSS when compared to normal care.


Quality of evidence

The quality of evidence would be considered moderate due to a serious risk of bias with regard to allocation and blinding in the trialsassessed in the reviews

Moderate



122 of 157

https://www.magicapp.org/app#/guideline/1242

Rationale

The QASC trial showed high quality evidence that monitoring of BGLs and treatment of hyperglycaemia > 11 mmol/L improves outcomesat 90 days, when used as part of a bundle of care.

The latest systematic reviews did not show any benefit for using IV insulin, however the individual trials included all showed some signs ofbias.


IV insulin infusion increases acuity of patient care, some general wards would not be able to care for a patient on an insulin infusion. Ec o n o m i c s W P t o r e v i e w .




Intervention: FeSS protocol

Comparator: No FeSS protocol

Summary

The Quality in Acute Stroke Care (QASC) trial reported by Middleton et al (2011) [89] was a cluster randomised trial (N = 1696) ofa treatment protocol (FeSS) for managing fever, glycaemia, and swallowing dysfunction. The trial showed high quality evidencethat monitoring of BGLs and treatment of hyperglycaemia > 11 mmol/L improves outcomes at 90 days, when used as part of abundle of care, although the effects of individual components of the intervention cannot be separated. Therefore, the evidence forthe benefits of hyperglycaemia management specifically is somewhat indirect.

Drury et al (2014) [98] provides a systematic evaluation of records and data that documents current stroke management practicesof the pre-intervention cohort prospectively recruited for the QASC trial. Retrospective medical record audits of all 19participating stroke units (n=718) revealed:

• 138 (19%) had four hourly or more temperature readings and 204 patients (29%) had a fever, with 44 (22%) receivingparacetamol.• A quarter of patients (n = 102/412, 25%) had six hourly or more glucose readings and 23% (95/412) had hyperglycemia,with 31% (29/95) of these treated with insulin.• The majority of patients received a swallow assessment (n = 562, 78%) by a speech pathologist in the first instance ratherthan a swallow screen by a nonspeech pathologist (n = 156, 22%). Of those who passed a screen (n = 108 of 156, 69%), 68% (n= 73) were reassessed by a speech pathologist and 97% (n = 71) were reconfirmed to be able to swallow safely.

Note: The statistical analysis used in Middleton et al (2011) [89] estimates absolute risk differences directly, and relative effectswere not really reported. The absolute differences entered are those reported in the study. The raw numbers of events in thecontrol group are used to calculate baseline risk, with the reported absolute risk difference then used to calculate risk in theintervention group. Relative effects have been left blank.

OutcomeTimeframe



No FeSS protocol FeSS protocol

Certainty ineffect


Summary

Death ordependency

90 days

n/a


577per 1000

420per 1000 High

Patients treated in strokecare units with FeSS

protocols have improveddeath or dependency


123 of 157

9 Critical



outcomes whencompared to patientstreated in stroke care

units without FeSSprotocols.

Functionaldependency

(Barthel Index>= 60)90 days

7 Critical

n/a



898per 1000

923per 1000


High

There is little or nodifference in functional

dependency as measuredby Barthel Index >= 60

for those treated instroke care units withFeSS protocols when

compared to thosetreated in stroke care




7 Critical

n/a



600per 1000

695per 1000

Difference: 95 more per 1000( CI 95% 50 fewer - 195 fewer )

High






Length of stay

7 Critical

Measured by: Length of StayLower better



Follow up 90 days

13.7days (Mean)

11.3days (Mean)

Difference: MD 1.5 fewer( CI 95% 0.5 more - 3.5 fewer )

High

There is no difference inmean length of stay forthose treated in stroke

care units with FeSSprotocols when




Death ordependency


Risk of bias: No seriousInconsistency: No seriousIndirectness: No serious Exclusion palliative patients may have under represented severestroke patients. ;Imprecision: No seriousPublication bias: No serious



Intervention reference:Primary studyBaseline/comparatorreference: Control arm of

Risk of bias: No seriousInconsistency: No seriousIndirectness: No serious Exclusion palliative patients may have resulted in severe strokesbeing under represented. ;


124 of 157

References

[98] Drury P, Levi C, McInnes E, Hardy J, Ward J, Grimshaw JM, D' Este C, Dale S, McElduff P, Cheung NW, Quinn C, Griffiths R,Evans M, Cadilhac D, Middleton S Management of fever, hyperglycemia, and swallowing dysfunction following hospital admissionfor acute stroke in New South Wales, Australia.. International journal of stroke : official journal of the International Stroke Society2014;9(1):23-31- Pubmed Journal






Risk of bias: No seriousInconsistency: No seriousIndirectness: No serious Exclusion palliative patients may have resulted in underrepresentation severe strokes. ;Imprecision: No seriousPublication bias: No serious

Length of stay




Population: Adults with stroke with hyperglycaemia

Intervention: Intravenous insulin

Comparator: Subcutaneous insulin

Summary

Two systematic reviews (Bellolio et al 2014 [85]; Ntaios et al 2014 [86]) were included. The Cochrane systematic review (Bellolioet al 2014) included 11 trials (N=1583 participants) and the other review (Ntaios et al 2014) included 9 trials (N=1491participants). Both reviews were consistent and reported no benefits from intensive therapy with IV insulin but also an increasedrate of complications (hypoglycemia). Early and intense therapy via IV insulin is not recommended.

OutcomeTimeframe



Subcutaneous insulin Intravenous insulin

Certainty ineffect


Summary

Mortality1,3 or 4 months

Odds Ratio 1.1(CI 95% 0.59 - 2.07)

115 125 ModerateIV insulin has little or nodifference on mortality


125 of 157


http://dx.doi.org/10.1111/ijs.12194

9 Critical


(Randomized controlled)Follow up 1,3 and 4

months

per 1000 per 1000


Due to serious riskof bias

when compared tosubcutaneous insulin

Functionaloutcome

90 days

9 Critical

Odds Ratio 1.2(CI 95% 0.8 - 1.8)


(Randomized controlled)Follow up 1, 3 & 4 months

414per 1000

459per 1000


ModerateDue to serious risk

of bias

IV insulin may have littleor no difference on

functional outcome whencompared to

subcutaneous insulin

Hypoglycaemia(asymptomatic

or symptomatic)

8 Critical

Odds Ratio 5.59(CI 95% 2.67 - 11.71)



11per 1000

59per 1000



of bias

Risk of hypoglycemia(asymptomatic or

symptomatic) isincreased with

intravenous insulincompared to


Practical issues Subcutaneous insulin Intravenous insulin Both

Procedure anddeviceIntravenous insulininfusion increasesacuity of patientcare.

Cannot be managed on mostgeneral wards requiring high

dependency bed (stroke unit orcritical care bed)


Mortality

Intervention reference:Systematic reviewBaseline/comparatorreference: Systematicreview [(86] with includedstudies:

Risk of bias: Serious Inadequate/lack of blinding of outcome assessors, resulting in potentialfor detection bias, Inadequate/lack of blinding of participants and personnel, resulting inpotential for performance bias, Inadequate concealment of allocation during randomizationprocess, resulting in potential for selection bias ;Inconsistency: No seriousIndirectness: No seriousImprecision: No serious

Functionaloutcome

Intervention reference:Systematic review [86]with included studies:Baseline/comparatorreference: Systematicreview [(86] with includedstudies:

Risk of bias: Serious Inadequate/lack of blinding of participants and personnel, resulting inpotential for performance bias, Inadequate/lack of blinding of outcome assessors, resultingin potential for detection bias, Inadequate concealment of allocation during randomizationprocess, resulting in potential for selection bias ;Inconsistency: No seriousIndirectness: No seriousImprecision: No seriousPublication bias: No serious


126 of 157


An early intensive approach to the maintenance of tight glycaemic control (between 4.0-7.5mmol/L) is not recommended. (Bellolio et al2014 [85]; Ntaios et al 2014 [86];

Practical Info

The trials in the Cochrane review used tight control of blood glucose (4 - 7.5mmol/L) the QASC trial (Middleton et al 2011) used suggestedthat insulin should only be used to maintain BGL of less than 11mmol/L (euglycaemia) as part of a care bundle.

The Australian Diabetes Society Guidelines for Routine Glucose Control in Hospital recommend:

1. Patients admitted to hospital with acute thrombotic stroke who have hyperglycaemia, should betreated to achieve and maintain glucose levels less than 10 mmol/L.2. Hypoglycaemia must be avoided, and therefore it would be prudent to avoid treatment which lowersthe glucose below 5 mmol/L.

Key Info

References

[85] Bellolio MF, Gilmore RM, Ganti L Insulin for glycaemic control in acute ischaemic stroke.. The Cochrane database ofsystematic reviews 2014;1 CD005346- Pubmed Journal

[86] Ntaios G, Papavasileiou V, Bargiota A, Makaritsis K, Michel P Intravenous insulin treatment in acute stroke: a systematicreview and meta-analysis of randomized controlled trials.. International journal of stroke : official journal of the InternationalStroke Society 2014;9(4):489-93- Pubmed Journal

Hypoglycaemia(asymptomatic or

symptomatic)


Risk of bias: Serious Inadequate/lack of blinding of participants and personnel, resulting inpotential for performance bias, Inadequate/lack of blinding of outcome assessors, resultingin potential for detection bias, Inadequate concealment of allocation during randomizationprocess, resulting in potential for selection bias ;Inconsistency: No seriousIndirectness: No seriousImprecision: No serious Low number of patients=pooled data ;Publication bias: No serious

Benefits and harms

The risk of hypoglycaemia was higher in the intervention groups treated with IV insulin (maintain glycaemic level between 4 - 7.5mmol/L) when compared with SC insulin or normal saline. This applied to symptomatic and non-symptomatic hypoglycaemia.

IV insulin has little or no difference on mortality or functional outcome when compared to SC insulin or normal saline.

IV insulin has little or no difference on dependency, death, indepence in daily activities or NIHSS when compared to normal care.



127 of 157





Rationale

Two systematic reviews (Bellolio et al 2014; Ntaios et al 2014) were included. The Cochrane systematic review (Bellolio 2014) included11 trials (N=1583 participants) and the other review (Ntaios 2014) included 9 trials (N=1491 participants). Both reviews were consistentand reported no benefits from intensive therapy with IV insulin but also an increase rate of complications (hypoglycemia). Early andintense therapy via IV insulin is not recommended.

Quality of evidence

The quality of evidence would be considered moderate due to a serious risk of bias with regard to allocation and blinding in the trialsassessed in the reviews

Moderate



IV insulin infusion increases acuity of patient care, some general wards would not be able to care for a patient on an insulin infusion. Ec o n o m i c s W P t o r e v i e w .






Summary

The Quality in Acute Stroke Care (QASC) trial reported by Middleton et al (2011) [89] was a cluster randomised trial (N = 1696) ofa treatment protocol (FeSS) for managing fever, glycaemia, and swallowing dysfunction. The trial showed high quality evidencethat monitoring of BGLs and treatment of hyperglycaemia > 11 mmol/L improves outcomes at 90 days, when used as part of abundle of care, although the effects of individual components of the intervention cannot be separated. Therefore, the evidence forthe benefits of hyperglycaemia management specifically is somewhat indirect.

Drury et al (2014) [98] provides a systematic evaluation of records and data that documents current stroke management practicesof the pre-intervention cohort prospectively recruited for the QASC trial. Retrospective medical record audits of all 19participating stroke units (n=718) revealed:


Note: The statistical analysis used in Middleton et al (2011) [89] estimates absolute risk differences directly, and relative effectswere not really reported. The absolute differences entered are those reported in the study. The raw numbers of events in thecontrol group are used to calculate baseline risk, with the reported absolute risk difference then used to calculate risk in theintervention group. Relative effects have been left blank.

OutcomeTimeframe




Certainty ineffect

estimatesSummary


128 of 157

(Quality ofevidence)

Death ordependency

90 days

9 Critical

n/a



577per 1000

420per 1000


High







7 Critical

n/a



898per 1000

923per 1000


High








7 Critical

n/a



600per 1000

695per 1000


High






Length of stay

7 Critical




Follow up 90 days

13.7days (Mean)

11.3days (Mean)


High






Death ordependency

Intervention reference:Primary studyBaseline/comparatorreference: Control arm of

Risk of bias: No seriousInconsistency: No seriousIndirectness: No serious Exclusion palliative patients may have under represented severestroke patients. ;


129 of 157

References







Risk of bias: No seriousInconsistency: No seriousIndirectness: No serious Exclusion palliative patients may have resulted in severe strokesbeing under represented. ;Imprecision: No seriousPublication bias: No serious





Length of stay






Comparator: Subcutaneous insulin

Summary



130 of 157



OutcomeTimeframe



Subcutaneous insulin Intravenous insulin

Certainty ineffect


Summary

Mortality1,3 or 4 months

9 Critical

Odds Ratio 1.1(CI 95% 0.59 - 2.07)


(Randomized controlled)Follow up 1,3 and 4

months

115per 1000

125per 1000



of bias

IV insulin has little or nodifference on mortality

when compared tosubcutaneous insulin

Functionaloutcome

90 days

9 Critical

Odds Ratio 1.2(CI 95% 0.8 - 1.8)



414per 1000

459per 1000



of bias

IV insulin may have littleor no difference on

functional outcome whencompared to


Hypoglycaemia(asymptomatic

or symptomatic)

8 Critical

Odds Ratio 5.59(CI 95% 2.67 - 11.71)



11per 1000

59per 1000



of bias

Risk of hypoglycemia(asymptomatic or

symptomatic) isincreased with

intravenous insulincompared to


Practical issues Subcutaneous insulin Intravenous insulin Both

Procedure anddeviceIntravenous insulininfusion increasesacuity of patientcare.

Cannot be managed on mostgeneral wards requiring high

dependency bed (stroke unit orcritical care bed)


Mortality


Risk of bias: Serious Inadequate/lack of blinding of outcome assessors, resulting in potentialfor detection bias, Inadequate/lack of blinding of participants and personnel, resulting inpotential for performance bias, Inadequate concealment of allocation during randomizationprocess, resulting in potential for selection bias ;Inconsistency: No seriousIndirectness: No seriousImprecision: No serious

Functionaloutcome

Intervention reference:Systematic review [86]with included studies:

Risk of bias: Serious Inadequate/lack of blinding of participants and personnel, resulting inpotential for performance bias, Inadequate/lack of blinding of outcome assessors, resultingin potential for detection bias, Inadequate concealment of allocation during randomization


131 of 157

References



Baseline/comparatorreference: Systematicreview [(86] with includedstudies:

process, resulting in potential for selection bias ;Inconsistency: No seriousIndirectness: No seriousImprecision: No seriousPublication bias: No serious

Hypoglycaemia(asymptomatic or

symptomatic)


Risk of bias: Serious Inadequate/lack of blinding of participants and personnel, resulting inpotential for performance bias, Inadequate/lack of blinding of outcome assessors, resultingin potential for detection bias, Inadequate concealment of allocation during randomizationprocess, resulting in potential for selection bias ;Inconsistency: No seriousIndirectness: No seriousImprecision: No serious Low number of patients=pooled data ;Publication bias: No serious




Comparator: Normal saline

Summary


OutcomeTimeframe



Normal saline Intravenous insulin

Certainty ineffect


Summary

Mortality30- 90 days

Relative risk 1.17(CI 95% 0.88 - 1.55)


258per 1000

302per 1000


of bias

Intravenous insulin haslittle or no difference on

mortality when comparedto normal saline


132 of 157





References


9 Critical

(Randomized controlled)Follow up 30-90 days, x1

unspecified


Functionaloutcome

30-90 days

7 Critical

Odds Ratio 0.9(CI 95% 0.68 - 1.19)



287per 1000

266per 1000



of bias

Intravenous insulin haslittle or no difference on

functional outcome whencompared to normal

saline

Asymptomaticor symptomatichypoglycaemia

30-90 days

7 Critical

Relative risk 9.4(CI 95% 6.03 - 14.65)


(Randomized controlled)Follow up 30-90 days

2per 1000

19per 1000



of bias

The risk of anyhypoglycaemia is

increased withintravenous insulin when

compared with nornalsaline but no effect on

symptomatichypoglycaemia rates.


Mortality

Intervention reference:Systematic review [86]with included studies:Baseline/comparatorreference: Systematicreview

Risk of bias: Serious Inadequate/lack of blinding of participants and personnel, resulting inpotential for performance bias, Inadequate/lack of blinding of outcome assessors, resultingin potential for detection bias ;Inconsistency: No seriousIndirectness: No seriousImprecision: No seriousPublication bias: No serious

Functionaloutcome



Asymptomatic orsymptomatic

hypoglycaemia




133 of 157






Intervention: Insulin for glycaemic control


Summary


OutcomeTimeframe



Usual care Insulin for glycaemiccontrol

Certainty ineffect


Summary

Dependency ordeath at the endof the follow-up

9 Critical

Odds Ratio 0.99(CI 95% 0.79 - 1.23)


(Randomized controlled)Follow up <30 days to 90

days

658per 1000

656per 1000



of bias

Insulin for glycaemiccontrol in acute

ischaemic strokeprobably has little or no

difference ondependency or death atthe end of the follow-up

Death

9 Critical

Odds Ratio 1.09(CI 95% 0.85 - 1.41)


(Randomized controlled)Follow up discharge-120

days

224per 1000

239per 1000



of bias



difference on death

Dependency ordeath - patients

with diabetesmellitus

9 Critical

Odds Ratio 0.66(CI 95% 0.35 - 1.24)


(Randomized controlled)Follow up 30-90 days

524per 1000

421per 1000



of bias



difference ondependency or death


134 of 157



among patients withdiabetes mellitus.


without diabetesmellitus

9 Critical

Odds Ratio 1.02(CI 95% 0.81 - 1.3)



676per 1000

680per 1000



of bias



difference ondependency or death

among patients withoutdiabetes mellitus


Dependency ordeath at the end of

the follow-up

Intervention reference:Systematic review [85]with included studies:Azevedo 2009, GIST-UK2007, GRASP 2009,INSULINFARCT 2012,Kreisel 2009, McCormick2010, Staszewski 2011,THIS 2008, Vinychuk2005,Baseline/comparatorreference: Control arm ofreference used forintervention

Risk of bias: Serious Most of the studies did not use blinded assessors, resulting in potentialfor performance bias. Inadequate concealment of allocation, in 5 studies, duringrandomization process, resulting in potential for selection bias, GIST-UK stopping earlierthan scheduled, due to slow enrollment rate resulting in potential for overestimatingbenefits. 4 studies had a high risk of bias secondary to inadequate allocation. ;Inconsistency: No seriousIndirectness: No seriousImprecision: No serious Smaller numbers in individual trials. ;Publication bias: No serious

Death

Intervention reference:Systematic review [85]with included studies:Azevedo 2009, GIST-UK2007, GRASP 2009,INSULINFARCT 2012,Kreisel 2009, McCormick2010, Staszewski 2011,THIS 2008, Walters 2006,Baseline/comparatorreference: Control arm ofreference used forintervention

Risk of bias: Serious Most of the studies did not use blinded assessors, resulting in potentialfor performance bias. Inadequate concealment of allocation, in 5 studies, duringrandomization process, resulting in potential for selection bias, GIST-UK stopping earlierthan scheduled, due to slow enrollment rate resulting in potential for overestimatingbenefits. 4 studies had a high risk of bias ;Inconsistency: No seriousIndirectness: No seriousImprecision: No seriousPublication bias: No serious


with diabetesmellitus

Intervention reference:Systematic review [85]with included studies:GRASP 2009, THIS 2008,Vinychuk 2005,Baseline/comparatorreference: Control arm ofreference used forintervention

Risk of bias: Serious Inadequate/lack of blinding of participants and personnel, resulting inpotential for performance bias for all three studies. Inadequate/lack of blinding of outcomeassessors, resulting in potential for detection bias in the Vinychuk 2005 study. ;Inconsistency: No seriousIndirectness: No seriousImprecision: No serious Low number of patients in each study but better numbers whenpooled. ;Publication bias: No serious


135 of 157

References



without diabetesmellitus

Intervention reference:Systematic review [85]with included studies:GIST-UK 2007,INSULINFARCT 2012,Kreisel 2009, McCormick2010, Staszewski 2011,Vinychuk 2005,Baseline/comparatorreference: Control arm ofreference used forintervention

Risk of bias: Serious Inadequate/lack of blinding of participants and personnel, resulting inpotential for performance bias for all three studies. Inadequate/lack of blinding of outcomeassessors, resulting in potential for detection bias in the Vinychuk 2005 study., Inadequate/lack of blinding of participants and personnel, resulting in potential for performance bias,Inadequate/lack of blinding of outcome assessors, resulting in potential for detection bias ;Inconsistency: No seriousIndirectness: No seriousImprecision: No seriousPublication bias: No serious Low number of patients in each study but better numbers whenpooled. ;





Summary


OutcomeTimeframe




Certainty ineffect


Summary

Independent indaily activities

8 Critical

Odds Ratio 1.03(CI 95% 0.81 - 1.32)



120 Days

317per 1000

323per 1000



of bias

Insulin for glycaemiccontrol probably has little

or no difference onindependence in daily

activities when comparedto usual care

Functionalneurological

Measured by: NIHSS andESS

(n/a) (n/a) ModerateInsulin for glycaemic

control probably has little


136 of 157



outcome at theend of the

follow-up -NIHSS or ESS

7 Critical

Lower betterBased on data from:1,432 patients in 8


Follow up discharge to120 Days


Due to serious riskof bias

or no difference onfunctional neurological

outcome at the end of thefollow-up


outcome -patients with

diabetesmellitus

7 Critical




Follow up 30 - 60 days

(n/a) (n/a)



of bias



outcome among patientswith diabetes mellitus


outcome -patients without

diabetesmellitus

7 Critical


Lower betterBased on data from:1,286 patients in 6


Follow up 5 days - 120Days

(n/a) (n/a)



of bias



outcome among patientswithout diabetes mellitus


Independent indaily activities

Intervention reference:Systematic review [85]with included studies:Azevedo 2009, GIST-UK2007, GRASP 2009,INSULINFARCT 2012,Kreisel 2009, McCormick2010, Staszewski 2011,THIS 2008, Vinychuk2005,Baseline/comparatorreference: Control arm ofreference used forintervention

Risk of bias: Serious Inadequate concealment of allocation during randomization process,resulting in potential for selection bias in one study and unclear in three of the studiesInadequate/lack of blinding of participants and personnel, resulting in potential forperformance bias in all studies Inadequate/lack of blinding of outcome assessors, resulting inpotential for detection bias in five studies ;Inconsistency: No seriousIndirectness: No seriousImprecision: No seriousPublication bias: No serious


outcome at the endof the follow-up -

NIHSS or ESS

Intervention reference:Systematic review [85]with included studies:GIST-UK 2007,INSULINFARCT 2012,Kreisel 2009, Staszewski2011, THIS 2008,Vinychuk 2005,

Risk of bias: Serious Inadequate sequence generation/ generation of comparable groups,resulting in potential for selection bias in only one study Inadequate concealment ofallocation during randomization process, resulting in potential for selection bias in twostudies and unclear in another one study. Inadequate/lack of blinding of participants andpersonnel, resulting in potential for performance bias in all studies Inadequate/lack ofblinding of outcome assessors, resulting in potential for detection bias in five of the studies,Incomplete data and/or large loss to follow up in four studies ;Inconsistency: No serious


137 of 157

References


Vriesendorp 2009, Walters2006,Baseline/comparatorreference: Control arm ofreference used forintervention



outcome - patientswith diabetes

mellitus

Intervention reference:Systematic review [85]with included studies:THIS 2008, Vinychuk2005, Walters 2006,Baseline/comparatorreference: Systematicreview

Risk of bias: Serious Lack of blinding of and personnel resulting in potential for performancebias in all three studies, only THIS 2008 reported blinding for participants. Lack of blinding ofoutcome assessors, resulting in potential for detection bias in two out of the three studies.Loss to follow up reported as none in Vinychuck 2005, however estimated to be at least 2.3%based on % of figures in follow up. ;Inconsistency: No seriousIndirectness: No seriousImprecision: No serious Low number of patients ;Publication bias: No serious


outcome - patientswithout diabetes

mellitus

Intervention reference:Systematic review [85]with included studies:GIST-UK 2007,INSULINFARCT 2012,Kreisel 2009, Staszewski2011, Vinychuk 2005,Vriesendorp 2009, GIST-UK 2007, INSULINFARCT2012, Kreisel 2009,Staszewski 2011, Vinychuk2005, Vriesendorp 2009,Baseline/comparatorreference: Control arm ofreference used forintervention

Risk of bias: Serious Vriesendorp 2009 used consecutive envelopes resulting in a potentialfor selection bias. Kriesel did not report information on concealment of allocation duringrandomization process, and Vriesendorp used consecutive envelopes resulting in potentialfor selection bias. All trials had a lack of blinding of participants and personnel, resulting inpotential for performance bias. Only GISTUK had adequate blinding of outcome assessors,resulting in potential for detection bias for the rest of the trials. GISTUK 2009 had a 7.4%loss to follow up and early termination. Vinychuk 2005 reported no loss to follow up butestimated to be 2.7% based on the figures reported. Vriesendorp 2009 havd 15% withoutoutcome documented at 5 Days. ;Inconsistency: No seriousIndirectness: No seriousImprecision: No seriousPublication bias: No serious






138 of 157



Summary


OutcomeTimeframe




Certainty ineffect


Summary

Symptomatichypoglycaemia

6 Important

Odds Ratio 14.6(CI 95% 6.62 - 32.21)


(Randomized controlled)Follow up 5 days - 120

days

4per 1000

55per 1000



of bias

This meta-analysisshowed a significant

difference in theincidence of

hypoglycaemia betweenthe treatment and

control groups suggestingthat insulin for glycaemiccontrol probably worsens

symptomatichypoglycaemia

Hypoglycaemia(with or without

symptoms)

6 Important

Odds Ratio 18.41(CI 95% 9.09 - 37.27)


(Randomized controlled)Follow up 5 days - 120

days

10per 1000

157per 1000



of bias

This meta-analysis founda significant difference in

the incidence ofhypoglycaemia between

the treatment andcontrol groups suggestingthat insulin for glycaemiccontrol probably worsens

hypoglycaemia (with orwithout symptoms.


Symptomatichypoglycaemia

Intervention reference:Systematic review [85]with included studies:Azevedo 2009, GIST-UK2007, GRASP 2009,INSULINFARCT 2012,Kreisel 2009, McCormick2010, Staszewski 2011,THIS 2008, Vriesendorp2009, Walters 2006,Baseline/comparatorreference: Control arm ofreference used forintervention

Risk of bias: Serious All trials had lack of blinding of personnel, and only THIS 2008 blindedthe participants resulting in potential for performance bias. Inadequate/lack of blinding ofoutcome assessors, resulting in potential for detection bias in six of the 10 trials. Two trialshad inadequate concealment of allocation during randomization process, and two it wasunclear int he reporting resulting in potential for selection bias. Two trials had incompletedata and/or large loss to follow up ;Inconsistency: No seriousIndirectness: No seriousImprecision: No seriousPublication bias: No serious


139 of 157

References


Hypoglycaemia(with or without

symptoms)

Intervention reference:Systematic review [85]with included studies:Azevedo 2009, GIST-UK2007, GRASP 2009,INSULINFARCT 2012,Kreisel 2009, McCormick2010, Staszewski 2011,THIS 2008, Vriesendorp2009, Walters 2006,Baseline/comparatorreference: Control arm ofreference used forintervention

Risk of bias: Serious Inadequate sequence generation/ generation of comparable groups,resulting in potential for selection bias in one trial, Inadequate/lack of blinding of participantsand personnel, resulting in potential for performance bias in all trials accept THIS 2008 whoblinded the participants. Only four trials had adequate/ of outcome assessors, resulting inpotential for detection bias. Incomplete data and/or large loss to follow up for two trials andunclear in two more. ;Inconsistency: No seriousIndirectness: No seriousImprecision: No seriousPublication bias: No serious


140 of 157



12 - Pyrexia management

In the initial period after a stroke, temperature higher than 37·5°C (Pyrexia) occurs in 20–50% of patients (Castillo et al 1999 [104]). Pyrexia is

associated with poorer outcomes after stroke (Greer et al 2008 [105]) and the most common causes of pyrexia are chest or urinary infections(Langhorne et al 2000 [106]). Fever in stroke patients needs to be managed proactively by the interdisciplinary team ideally as part of abundled care package where it has been demonstrated to reduce mortality and morbidity (Middelton et al 2011 [109]).


All patients should have their temperature monitored at least four times a day for 72 hours. (Middleton et al 2011 [97])

Practical Info

To avoid fatigue it was considered reasonable to undertake four observations over a 24 hour period rather than strict 6 hourly protocol.

Key Info

Rationale

Fever is an important indicator of developing sepsis which requires specific investigation and treatment. The QASC trial showed high

quality evidence that monitoring and treatment of fever > 37.5 oC improves outcomes at 90 days, when used as part of a bundle of care.The absolute benefits reported in this trial clearly outweigh the drawbacks/harms associated with not receiving this aspect of the carebundle (temperature recorded four times a day).

Benefits and harms

There were no harms reported in the patients who were treated in stroke units that had implemented the FeSS treatment protocols inthe QASC study. The substantial benefits of this care package (patients are 16% more likely to be alive and independent) compared tothose cared for in stroke units without FESS protocols warrant the recommendation that stroke units should follow similar protocols.

Therapeutic hypothermia probably has little or no difference on death or disability and may increase length of stay but furtherresearch is needed.


Quality of evidence

For the comparator FeSS vs no FeSS the quality of evidence is very high, as the evidence is from a large single blind RCT with minimalbias.Drury et al provides a systematic evaluation of records and data that documents current stroke management practices indicating theneed for urgent behaviour change.Therapeutic hypothermia is based on small pilot trials at risk of bias.

Moderate


There are no perceived risks or inconvenience to having temperature recorded four times a day within the first 72hrs. The low qualityevidence available for therapeutic hypothermia do not warrant consideration at this stage due to safety concerns related to seriouscomplication rates.



The only resource consideration would be the increased cost if use of Paracetamol infusions is required as part of the FeSS treatmentsprotocols.Although the reduced length of stay of two days was not statistically significant, the potential cost benefits warrant furtherconsideration. The small study exploring therapeutic hypothermia found this treatment probably increases length of stay. E c o n o mi c s W P t o r e v i e w .



141 of 157





Summary

The Quality in Acute Stroke Care (QASC) trial reported by Middleton et al (2011) [89] was a cluster randomised trial (N = 1696) ofa treatment protocol (FeSS) for managing fever, glycaemia, and swallowing dysfunction. The trial showed high quality evidence

that monitoring and treatment of fever > 37.5 oC improves outcomes at 90 days, when used as part of a bundle of care, althoughthe effects of individual components of the intervention cannot be separated. Therefore, the evidence for the benefits of pyrexiamanagement specifically is somewhat indirect.

Drury et al (2014) [98] provides a systematic evaluation of records and data that documents current stroke management practicesof the pre-intervention cohort prospectively recruited for the Quality in Acute Stroke Care trial. Retrospective medical recordaudits of all 19 participating stroke units (n=718) revealed:


Note: The statistical analysis used in Middleton et al (2011) [89] estimates absolute risk differences directly, and relative effectswere therefore not reported. The absolute differences entered are those reported in the study. The raw numbers of events in thecontrol group are used to calculate baseline risk, with the reported absolute risk difference then used to calculate (estimated) riskin the intervention group. Relative effects have been left blank.

OutcomeTimeframe




Certainty ineffect


Summary

Death ordependency

90 days

9 Critical

n/a



577per 1000

420per 1000


High






n/a

Based on data from 955898 923 High


dependency as measured


142 of 157


7 Critical


Follow up 90 days

per 1000 per 1000


by Barthel Index >= 60for those treated in

stroke care units withFeSS protocols when





7 Critical

n/a



600per 1000

695per 1000


High






Length of stay

7 Critical




Follow up 90 days

13.7days (Mean)

11.3days (Mean)


High






Death ordependency






Risk of bias: No seriousInconsistency: No seriousIndirectness: No serious Exclusion palliative patients may have resulted in severe strokesbeing under represented. ;Imprecision: No seriousPublication bias: No serious






143 of 157

Weak Recommendation

Antipyretic therapy, comprising paracetamol, may be used where fever > 37.5 o C. (den Hertog et al 2009 [102]; Middleton et al 2011 [97])

Practical Info

The PAIS trial evaluated use of high dose Paracetamol (6g per day x3 days) for those with temperature b/w 36-39 degrees celcius.Exclusions included any trial participants with liver dysfunction or alcohol abuse, and pre morbid mRS > 2.

Key Info

References

[97] Middleton S, McElduff P, Ward J, Grimshaw JM, Dale S, D'Este C, Drury P, Griffiths R, Cheung NW, Quinn C, Evans M,Cadilhac D, Levi C Implementation of evidence-based treatment protocols to manage fever, hyperglycaemia, and swallowingdysfunction in acute stroke (QASC): a cluster randomised controlled trial.. Lancet (London, England) 2011;378(9804):1699-706-Pubmed Journal


Length of stay



Benefits and harms

There was no evidence from 1 RCT that routine use of paracetamol improved outcome but in a post-hoc subgroup those with fever>37.5 appeared to benefit. There were no harms reported in the patients who were treated in stroke units that had implemented theFeSS treatment protocols. The substantial benefits (patients are 16% more likely to be alive and independent) compared to thosecared for in stroke units without FeSS treatment protocols warrant the recommendation that stroke units should follow similarprotocols. The FeSS treatment protocols reccommend treatment with paracetamol for temp >37.5 degrees celcius.

Therapeutic hypothermia probably has little or no difference on death or disability and may increase length of stay but furtherresearch is needed.

Results from PAIS trial do not support routine use of high-dose paracetamol for all acute stroke patients but some benefit observedfor those patients with temperature 37-39 degrees celcius.No statistical significance although improvement was noted in disability (mRS) for those treated with paracetamol, no adverse eventsor harms were observed in the treatment group or comparator.


Quality of evidence Moderate


144 of 157


http://dx.doi.org/10.1016/S0140-6736(11)61485-2



Rationale

Fever is an important indicator of developing sepsis which requires specific investigation and treatment. The QASC trial showed high

quality evidence that monitoring and treatment of fever > 37.5 oC improves outcomes at 90 days, when used as part of a bundle of care.Results from PAIS trial do not support routine use of high-dose paracetamol for all acute stroke patients but some benefit observed forthose patients with temperature 37-39 degrees celcius although this subgroup not defined in advance.The PAIS trial showed no difference between treatment groups. They also updated the meta-analyses (Cochrane 2009) using these datawhich showed no significant difference when comparing active treatment and control therefore routine paracetamol is not recommended.

For the comparator FeSS vs no FeSS the quality of evidence is very high, as the evidence is from a large single blind RCT with minimalbias.The Drury et al provides a systematic evaluation of records and data that documents current stroke management practices indicatingthe need for urgent behaviour change.Therapeutic hypothermia is based on small pilot trials at risk of bias.PAIS study (Paracetamol vs placebo) was of high quality however the analysis plan was changed during the study from fixeddichotomous mRS to sliding dichotomous analysis for mRS. There was no difference between the groups using either tool.



The only resource consideration would be the increased cost if use of Paracetamol infusions is required as part of the FeSS treatmentsprotocols.Although the reduced length of stay of two days was not statistically significant, the potential cost benefits warrant furtherconsideration. The small study exploring therapeutic hypothermia found this treatment probably increases length of stay.




Intervention: Paracetamol

Comparator: Placebo

Summary

A Cochrane review (Den Hertog et al 2009 [103]) included five pharmacological temperature reduction trials and three physicalcooling trials (total of 423 participants). No benefits were found for either strategy in terms of reducing the risk of death ordependency (odds ratio (OR) 0.9, 95% confidence interval (CI) 0.6 to 1.4) or death (OR 0.9, 95% CI 0.5 to 1.5).

One large subsequent trial (Den Hertog et al 2009 [102]) including 1400 patients found no benefits for routine high doseparacetamol but some groups (such as those with fever) may benefit based on subgroup analysis. An updated meta-analysisincluding this trial as well as the trials in the earlier review showed no significant increase in favourable outcomes.

Note: Den Hertog et al (2009) [102] reported both unadjusted and covariate-adjusted odds ratios for the disability outcome. Thecovariate-adjusted value is reported here.

OutcomeTimeframe



Placebo Paracetamol

Certainty ineffect


Summary


145 of 157

Practice Statement

Antipyretics can be given orally or via a nasogastric tube or suppository (for those with dysphagia)

References

[102] den Hertog HM, van der Worp HB, van Gemert HMA, Algra A, Kappelle LJ, van Gijn J, Koudstaal PJ, Dippel DWJ TheParacetamol (Acetaminophen) In Stroke (PAIS) trial: a multicentre, randomised, placebo-controlled, phase III trial.. The Lancet.Neurology 2009;8(5):434-40- Pubmed Journal

Disability:favourable

outcome (mRS<= 2)

9 Critical

Odds Ratio 1.02(CI 95% 0.78 - 1.32)



500per 1000

505per 1000



imprecision

More patients in theparacetamol group than

in the placebo groupimproved beyond

expectation, but this wasnot statistically

significant thereforeparacetamol has little or

no difference on disability(mRS) when compared

with placebo


8 Critical

n/a


(Randomized controlled)Follow up Discharge

100per 1000

80per 1000

Difference: 20 fewer per 1000CI 95%

HighParacetamol has little orno difference on serious

adverse events


Disability:favourable

outcome (mRS <=2)


Risk of bias: No seriousInconsistency: No seriousIndirectness: No serious The analysis plan for PAIS was changed from a fixed dichotomy ofthe mRS to the sliding dichotomy analysis during the trial, neither showed an effect ofparacetamol on functional outcome. ;Imprecision: Serious Wide confidence intervals ;Publication bias: No serious





146 of 157


http://dx.doi.org/10.1016/S1474-4422(09)70051-1

13 - Dysphagia

Dysphagia is a common consequence of acute stroke with a reported incidence of 47% in the most recent national audit.11 Dysphagia isassociated with an increased risk of complications, such as aspiration pneumonia, dehydration and malnutrition.479, 489 Dysphagia was alsofound to lead to poor clinical outcomes (chest infection, death, disability, discharge destination, longer LOS) reinforcing the need for earlydetection and management.490Adherence to a formal dysphagia screening protocol reduces the incidence of pneumonia in acute stroke patients.491, 492 Another studyimplementing evidence-based acute care involving dysphagia screening, referral and assessment demonstrated improved process and patientoutcomes.493 Further studies are needed to clarify the key elements that improve outcomes, including identifying which screening tool ismost useful.Several systematic reviews agree on the value of early screening using bedside tools.494, 495 Due to variability in the studies, threesystematic reviews were unable to conclude which screening tool was most useful.495-497 While most tests had sensitivities of 70–90%some were much lower, the lowest reported being 42%. Specificity ranged from 22% to 67% in one review 496 and 59% to 91% in another.495Subsequent studies of bedside clinical screening have demonstrated similar sensitivities with other bedside tests.498-504 Two recent well-developed and validated tests include the Gugging Swallowing Screen (GSS) 502 and the Toronto Bedside Swallowing Screening test (TOR-BSST).503 The GSS involves indirect and then direct swallowing tests. .Accuracy was good: sensitivity 100%, specificity 50–69%, PPV74–81%, NPV 100%, area under ROC curve 0.933. Inter-rater reliability was excellent.502 The TOR-BSST was designed to be used by anyprofessional trained in assessment of stroke across all settings and includes five items: Kidd water swallow test, pharyngeal sensation, tonguemovement and general dysphonia, voice before and voice after. Overall (acute and rehabilitation phases) accuracy (sensitivity 91.3%,specificity 66.7%, PPV 50–77%, NPV 90–93%) was similar to other tests. Inter-rater reliability was also very good (ICC 0.92).503 Thecombination of a bedside screening test and monitoring of oxygen saturations improves the sensitivity of earlier bedside tests(87–100%).498, 500, 505 The gag reflex was not found to be a valid screen for dysphagia and should therefore not be used.496, 497Screening tools have been developed for use by non-specialist staff who must undertake essential training prior to using such tools.496Ideally the initial screen would be undertaken by a speech pathologist as part of a comprehensive assessment. However, it is not feasible tooffer such a service 24 hours a day, seven days a week, hence consideration needs to be given to resource and training requirements forestablishing and maintaining effective dysphagia screening. Additional resources may need to be considered for initial and ongoing training(particularly in view of the high staff turnover in some ED departments) and the development of local protocols for ensuring routineimplementation of swallow screening (including rostering that ensures appropriately trained staff are available on all shifts).Videofluoroscopic modified barium swallow (VMBS) study is considered the reference standard to confirm swallowing dysfunction andpresence of aspiration. Factors limiting its use include: the relatively complex, time-consuming and resource- intensive nature of the test; thesmall exposure to radiation; and the difficulty of positioning patients appropriately. In addition, the results can be difficult to interpret andvariation among individual raters may occur.495 There is no agreed criterion for when a VMBS study is required and local policies should bedeveloped that take into consideration local resources and the potential limitations noted above.Fibre-optic endoscopic evaluation of swallowing (FEES) has also been used as a reference standard in studies assessing dysphagia screeningtools 498-501 and has been found to have similar sensitivity and specificity to VMBS.506 FEES is portable (possibly allowing more immediateaccess and saving time), requires less staff and is therefore cheaper, and reduces radiation exposure.506 FEES is generally well tolerated but isassociated with a small increase in nose bleeds (6%) and adverse effects on SBP, HR and oxygen saturation (although not severe).507 Whilespeech pathologists currently coordinate and conduct VMBS studies, only specialists with recognised training and credentials can conductFEES and it is therefore not commonly available in Australia.Strategies to prevent complications and restore the normal swallow have been described as either direct/compensatory (such as fluid and dietmodification, safe swallowing strategies and optimising the position of the stroke survivor while eating) or indirect (such as oral musculatureexercises and stimulation of the oral and pharyngeal structures).508 Discussion about the intensity of interventions is included in 6.1 Amountand timing of rehabilitation.A systematic review (15 RCTs) of a range of interventions concluded there was general support for dysphagia interventions but as few RCTsused the same intervention or outcomes the interpretation of the evidence was limited.509 Two RCTs within this review found compensatoryand intervention-swallowing techniques in combination with texture-modified diets can increase safe swallowing. Due to significantheterogeneity, no conclusions could be made about the effect of dietary texture modifications and/or alteration of fluid viscosity (fourtrials).509 One subsequent RCT found spoon-thick consistency reduced the risk of aspiration compared with fluid consistency (RR 0.13, 95%CI 0.04–0.39).504One Cochrane review found insufficient evidence (one trial, n=66) to determine the effects of acupuncture on dysphagia.510In a number of small trials, both neuromuscular electrical stimulation (NMES) and thermal tactile stimulation (TTS) reduced the severity ofswallow impairment.511-513 In one small subsequent small RCT (n=25), no difference between NMES and conventional therapy was found.Both interventions resulted in measurable improvement according to patient’s perception, nutrition and oral motor function test but notaccording to videoradiographic findings suggesting patient perception of improvement in swallowing may be erroneous.514 Another smallquasi-RCT (n=28) found sensory stimulation plus TTS is better than TTS alone.515 Electrical stimulation remains an evolving area ofdysphagia treatment. Possible contraindications must be assessed (e.g. pregnancy, presence of a pacemaker). Electrical stimulation shouldonly be considered by clinicians experienced with this intervention and applied according to published parameters. Further information canbe obtained from Speech Pathology Australia (www.speechpathologyaustralia.org.au/library/Neuromuscular_Electrical_Stimulation_NMES_Position_Statement.pdf).Therapy targeting specific muscle groups (e.g. ‘Shaker’ therapy) appears beneficial for people with specific dysphagia (two small RCTs (n=27,n=9).516, 517 Another small RCT (n=26) found that repetitive transcranial magnetic stimulation (rTMS) in addition to usual care may improvefunctional swallowing as measured by bedside assessment).518


147 of 157

Dysphagia usually improves within a few weeks following stroke; however, it can persist, requiring long-term intervention and/or alternativefeeding strategies (see 7.1 Hydration and nutrition). Patients with significant dysphagia who are unable to manage their secretions sometimesundergo a tracheostomy. Management of such patients should incorporate relevant local and international protocols regarding tracheostomyas well as the specific dysphagia interventions outlined below.


148 of 157

References

[1] Stroke Unit Trialists' Collaboration Organised inpatient (stroke unit) care for stroke. Cochrane Database of Systematic Reviews 2013;(9):Journal

[2] Sun Y, Paulus D, Eyssen M, Maervoet J, Saka O A systematic review and meta-analysis of acute stroke unit care: what's beyond thestatistical significance?. BMC medical research methodology 2013;13 132- Pubmed Journal

[3] Chan DK, Cordato D., O'Rourke F., Chan DL, Pollack M., Middleton S., Levi C. Comprehensive stroke units: a review of comparativeevidence and experience. International Journal of Stroke 2013;8(4):260-4 Journal

[4] Ciccone A, Celani MG, Chiaramonte R, Rossi C, Righetti E Continuous versus intermittent physiological monitoring for acute stroke.Cochrane Database of Systematic Reviews 2013;(2): Journal

[5] Moodie M, Cadilhac D, Pearce D, Mihalopoulos C, Carter R, Davis S, Donnan G Economic evaluation of Australian stroke services: aprospective, multicenter study comparing dedicated stroke units with other care modalities.. Stroke; a journal of cerebral circulation2006;37(11):2790-5- Pubmed

[6] Hunter RM, Davie C, Rudd A, Thompson A, Walker H, Thomson N, Mountford J, Schwamm L, Deanfield J, Thompson K, Dewan B, Mistry M,Quoraishi S, Morris S Impact on clinical and cost outcomes of a centralized approach to acute stroke care in London: a comparativeeffectiveness before and after model.. PloS one 2013;8(8):e70420- Pubmed Journal

[7] Te Ao BJ, Brown PM, Feigin VL, Anderson CS Are stroke units cost effective? Evidence from a New Zealand stroke incidence andpopulation-based study.. International journal of stroke : official journal of the International Stroke Society 2012;7(8):623-30- PubmedJournal

[8] National Stroke Foundation National Acute Stroke Services Framework. 2015;

[9] Gade G, Venohr I, Conner D, McGrady K, Beane J, Richardson RH, Williams MP, Liberson M, Blum M, Della Penna R Impact of an inpatientpalliative care team: a randomized control trial.. Journal of palliative medicine 2008;11(2):180-90- Pubmed Journal

[10] Eastman P, McCarthy G, Brand CA, Weir L, Gorelik A, Le B Who, why and when: stroke care unit patients seen by a palliative care servicewithin a large metropolitan teaching hospital.. BMJ supportive & palliative care 2013;3(1):77-83 Journal

[11] Gardiner C, Harrison M, Ryan T, Jones A Provision of palliative and end-of-life care in stroke units: a qualitative study.. Palliative medicine2013;27(9):855-60- Pubmed Journal

[12] Creutzfeldt CJ, Holloway RG, Walker M. Symptomatic and palliative care for stroke survivors. Journal of General Internal Medicine2012;27(7):853-860 Journal

[13] Burton CR, Payne S. Integrating palliative care within acute stroke services: Developing a programme theory of patient and family needs,preferences and staff perspectives. BMC Palliative Care 2012;11(22): Journal

[14] Payne S., Burton C., Addington-Hall J., Jones A. End-of-life issues in acute stroke care: a qualitative study of the experiences andpreferences of patients and families. Palliative Medicine 2010;24(2):146-53 Journal


149 of 157



http://dx.doi.org/10.1186/1471-2288-13-132

http://dx.doi.org/http://dx.doi.org/10.1111/j.1747-4949.2012.00850.x






http://dx.doi.org/10.1111/j.1747-4949.2011.00632.x


http://dx.doi.org/10.1089/jpm.2007.0055

http://dx.doi.org/10.1136/bmjspcare-2012-000331


http://dx.doi.org/10.1177/0269216313483846

http://dx.doi.org/http://dx.doi.org/10.1007/s11606-011-1966-4

http://dx.doi.org/http://dx.doi.org/10.1186/1472-684X-11-22

http://dx.doi.org/10.1177/0269216309350252

[15] de Boer ME, Depla M, Wojtkowiak J, Visser MC, Widdershoven GAM, Francke AL, Hertogh CMPM Life-and-death decision-making in theacute phase after a severe stroke: Interviews with relatives.. Palliative medicine 2015;29(5):451-7- Pubmed Journal

[16] Blacquiere D, Bhimji K, Meggison H, Sinclair J, Sharma M Satisfaction with palliative care after stroke: a prospective cohort study.. Stroke;a journal of cerebral circulation 2013;44(9):2617-9- Pubmed Journal

[17] Stevens T, Payne SA, Burton C, Addington-Hall J, Jones A Palliative care in stroke: a critical review of the literature.. Palliative medicine2007;21(4):323-31- Pubmed

[18] Jack C, Jones L, Jack BA, Gambles M, Murphy D, Ellershaw JE Towards a good death: the impact of the care of the dying pathway in anacute stroke unit.. Age and ageing 2004;33(6):625-6- Pubmed

[19] Thrift AG, Dewey HM, Macdonell RA, McNeil JJ, Donnan GA Stroke incidence on the east coast of Australia: the North East MelbourneStroke Incidence Study (NEMESIS).. Stroke; a journal of cerebral circulation 2000;31(9):2087-92- Pubmed


[21] Emberson J, Lees KR, Lyden P, Blackwell L, Albers G, Bluhmki E, Brott T, Cohen G, Davis S, Donnan G, Grotta J, Howard G, Kaste M, KogaM, von Kummer R, Lansberg M, Lindley RI, Murray G, Olivot JM, Parsons M, Tilley B, Toni D, Toyoda K, Wahlgren N, Wardlaw J, Whiteley W,del Zoppo GJ, Baigent C, Sandercock P, Hacke W Effect of treatment delay, age, and stroke severity on the effects of intravenous thrombolysiswith alteplase for acute ischaemic stroke: a meta-analysis of individual patient data from randomised trials.. Lancet (London, England)2014;384(9958):1929-35- Pubmed Journal

[22] The IST-3 collaborative group Effect of thrombolysis with alteplase within 6 h of acute ischaemic stroke on long-term outcomes (the thirdInternational Stroke Trial [IST-3]): 18-month follow-up of a randomised controlled trial.. The Lancet. Neurology 2013;12(8):768-76- PubmedJournal

[23] Anderson CS, Robinson T, Lindley RI, Arima H, Lavados PM, Lee T-H, Broderick JP, Chen X, Chen G, Sharma VK, Kim JS, Thang NH, Cao Y,Parsons MW, Levi C, Huang Y, Olavarría VV, Demchuk AM, Bath PM, Donnan GA, Martins S, Pontes-Neto OM, Silva F, Ricci S, Roffe C, PandianJ, Billot L, Woodward M, Li Q, Wang X, Wang J, Chalmers J Low-Dose versus Standard-Dose Intravenous Alteplase in Acute Ischemic Stroke. NEngl J Med 2016; Journal Website

[24] Wardlaw JM, Koumellis P, Liu M Thrombolysis (different doses, routes of administration and agents) for acute ischaemic stroke.. TheCochrane database of systematic reviews 2013;(5):CD000514- Pubmed Journal

[25] Saqqur M, Tsivgoulis G, Nicoli F, Skoloudik D, Sharma VK, Larrue V, Eggers J, Perren F, Charalampidis P, Storie D, Shuaib A, Alexandrov AVThe role of sonolysis and sonothrombolysis in acute ischemic stroke: a systematic review and meta-analysis of randomized controlled trialsand case-control studies.. Journal of neuroimaging : official journal of the American Society of Neuroimaging 2014;24(3):209-20- PubmedJournal

[26] Ricci S, Dinia L, Del Sette M, Anzola P, Mazzoli T, Cenciarelli S, Gandolfo C Sonothrombolysis for acute ischaemic stroke.. The Cochranedatabase of systematic reviews 2012;(6):CD008348- Pubmed Journal

[27] Demaerschalk BM, Hwang H-M, Leung G Cost analysis review of stroke centers, telestroke, and rt-PA.. The American journal of managedcare 2010;16(7):537-44- Pubmed

[28] Tan Tanny SP, Busija L, Liew D, Teo S, Davis SM, Yan B Cost-effectiveness of thrombolysis within 4.5 hours of acute ischemic stroke:experience from Australian stroke center.. Stroke; a journal of cerebral circulation 2013;44(8):2269-74- Pubmed Journal


150 of 157


http://dx.doi.org/10.1177/0269216314563427








http://dx.doi.org/10.1016/S0140-6736(14)60584-5


http://dx.doi.org/10.1016/S1474-4422(13)70130-3






http://dx.doi.org/10.1111/jon.12026






[29] Pan Y, Chen Q, Zhao X, Liao X, Wang C, Du W, Liu G, Liu L, Wang C, Wang Y, Wang Y, Cost-effectiveness of thrombolysis within 4.5 hoursof acute ischemic stroke in China.. PloS one 2014;9(10):e110525-null Pubmed Journal

[30] Boudreau DM, Guzauskas GF, Chen ER, Lalla D, Tayama D, Fagan SC, Veenstra DL Cost-effectiveness of recombinant tissue-typeplasminogen activator within 3 hours of acute ischemic stroke: current evidence.. Stroke; a journal of cerebral circulation2014;45(10):3032-9-null Pubmed Journal

[31] Tung CE, Win SS, Lansberg MG Cost-effectiveness of tissue-type plasminogen activator in the 3- to 4.5-hour time window for acuteischemic stroke.. Stroke; a journal of cerebral circulation 2011;42(8):2257-62-null Pubmed Journal

[32] Boudreau DM, Guzauskas G, Villa KF, Fagan SC, Veenstra DL A model of cost-effectiveness of tissue plasminogen activator in patientsubgroups 3 to 4.5 hours after onset of acute ischemic stroke.. Annals of emergency medicine 2013;61(1):46-55-null Pubmed Journal

[33] Coutts SB, Modi J, Patel SK, Demchuk AM, Goyal M, Hill MD, CT/CT angiography and MRI findings predict recurrent stroke aftertransient ischemic attack and minor stroke: results of the prospective CATCH study.. Stroke; a journal of cerebral circulation2012;43(4):1013-7-null Pubmed Journal

[34] Demaerschalk BM, Kleindorfer DO, Adeoye OM, Demchuk AM, Fugate JE, Grotta JC, Khalessi AA, Levy EI, Palesch YY, Prabhakaran S,Saposnik G, Saver JL, Smith EE, Scientific Rationale for the Inclusion and Exclusion Criteria for Intravenous Alteplase in Acute Ischemic Stroke:A Statement for Healthcare Professionals From the American Heart Association/American Stroke Association.. Stroke; a journal of cerebralcirculation 2016;47(2):581-641-null Pubmed Journal

[35] Badhiwala JH, Nassiri F, Alhazzani W, Selim MH, Farrokhyar F, Spears J, Kulkarni AV, Singh S, Alqahtani A, Rochwerg B, Alshahrani M,Murty NK, Alhazzani A, Yarascavitch B, Reddy K, Zaidat OO, Almenawer SA Endovascular Thrombectomy for Acute Ischemic Stroke: A Meta-analysis.. JAMA 2015;314(17):1832-43- Pubmed Journal

[36] Brinjikji W, Murad MH, Rabinstein AA, Cloft HJ, Lanzino G, Kallmes DF Conscious sedation versus general anesthesia duringendovascular acute ischemic stroke treatment: a systematic review and meta-analysis.. AJNR. American journal of neuroradiology2015;36(3):525-9- Pubmed Journal

[37] Campbell BC, Hill MD, Rubiera M, Menon BK, Demchuk A, Donnan GA, Roy D, Thornton J, Dorado L, Bonafe A, Levy EI, Diener H-C,Hernández-Pérez M, Pereira VM, Blasco J, Quesada H, Rempel J, Jahan R, Davis SM, Stouch BC, Mitchell PJ, Jovin TG, Saver JL, Goyal MSafety and Efficacy of Solitaire Stent Thrombectomy: Individual Patient Data Meta-Analysis of Randomized Trials . Stroke 2016/02/17 ;Website

[38] Goyal M, Menon BK, van Zwam WH, Dippel DWJ, Mitchell PJ, Demchuk AM, Dávalos A, Majoie CBLM, van der Lugt A, de Miquel MA,Donnan GA, Roos YBWEM, Bonafe A, Jahan R, Diener H-C, van den Berg LA, Levy EI, Berkhemer OA, Pereira VM, Rempel J, Millán M, DavisSM, Roy D, Thornton J, Román LS, Ribó M, Beumer D, Stouch B, Brown S, Campbell BCV, van Oostenbrugge RJ, Saver JL, Hill MD, Jovin TGEndovascular thrombectomy after large-vessel ischaemic stroke: a meta-analysis of individual patient data from five randomised trials..Lancet (London, England) 2016;387(10029):1723-31- Pubmed Journal

[39] Kim AS, Nguyen-Huynh M, Johnston SC A cost-utility analysis of mechanical thrombectomy as an adjunct to intravenous tissue-typeplasminogen activator for acute large-vessel ischemic stroke.. Stroke; a journal of cerebral circulation 2011;42(7):2013-8- Pubmed Journal

[40] Palesch YY, Yeatts SD, Tomsick TA, Foster LD, Demchuk AM, Khatri P, Hill MD, Jauch EC, Jovin TG, Yan B, von Kummer R, Molina CA,Goyal M, Schonewille WJ, Mazighi M, Engelter ST, Anderson C, Spilker J, Carrozzella J, Ryckborst KJ, Janis LS, Simpson A, Simpson KN,Broderick JP Twelve-Month Clinical and Quality-of-Life Outcomes in the Interventional Management of Stroke III Trial.. Stroke; a journal ofcerebral circulation 2015;46(5):1321-7- Pubmed Journal

[41] Puñal-Riobóo J, Atienza G, Blanco M Safety and Efficacy of Mechanical Thrombectomy Using Stent Retrievers in the EndovascularTreatment of Acute Ischaemic Stroke: A Systematic Review. Interventional Neurology 2015/06/03;3(3-4):149-164 Journal Website


151 of 157








http://dx.doi.org/10.1016/j.annemergmed.2012.04.020




http://dx.doi.org/10.1161/STR.0000000000000086


http://dx.doi.org/10.1001/jama.2015.13767


http://dx.doi.org/10.3174/ajnr.A4159



http://dx.doi.org/10.1016/S0140-6736(16)00163-X





http://dx.doi.org/10.1159/000430474

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4521200/

[42] Turk AS, Turner R, Spiotta A, Vargas J, Holmstedt C, Ozark S, Chalela J, Turan T, Adams R, Jauch EC, Battenhouse H, Whitsitt B, Wain M,Chaudry MI Comparison of endovascular treatment approaches for acute ischemic stroke: cost effectiveness, technical success, and clinicaloutcomes.. Journal of neurointerventional surgery 2015;7(9):666-70- Pubmed Journal

[43] Aronsson M, Persson J, Blomstrand C, Wester P, Levin L-Å Cost-effectiveness of endovascular thrombectomy in patients with acuteischemic stroke.. Neurology 2016;86(11):1053-9-null Pubmed Journal

[44] Ganesalingam J, Pizzo E, Morris S, Sunderland T, Ames D, Lobotesis K Cost-Utility Analysis of Mechanical Thrombectomy Using StentRetrievers in Acute Ischemic Stroke.. Stroke; a journal of cerebral circulation 2015;46(9):2591-8-null Pubmed Journal

[45] Lobotesis K, Veltkamp R, Carpenter IH, Claxton LM, Saver JL, Hodgson R Cost-effectiveness of stent-retriever thrombectomy incombination with IV t-PA compared with IV t-PA alone for acute ischemic stroke in the UK.. Journal of medical economics2016;19(8):785-94-null Pubmed Journal

[46] Rothwell PM, Algra A, Chen Z, Diener H-C, Norrving BO, Mehta Z Effects of aspirin on risk and severity of early recurrent stroke aftertransient ischaemic attack and ischaemic stroke: time-course analysis of randomised trials. The Lancet 2016; Journal Website

[47] Sandercock PAG, Counsell C, Kane EJ Anticoagulants for acute ischaemic stroke.. The Cochrane database of systematic reviews 2015;3CD000024- Pubmed Journal

[48] He F, Xia C, Zhang J-H, Li X-Q, Zhou Z-H, Li F-P, Li W, Lv Y, Chen H-S Clopidogrel plus aspirin versus aspirin alone for preventing earlyneurological deterioration in patients with acute ischemic stroke.. Journal of clinical neuroscience : official journal of the NeurosurgicalSociety of Australasia 2015;22(1):83-6- Pubmed Journal

[49] Yi X, Lin J, Wang C, Zhang B, Chi W A comparative study of dual versus monoantiplatelet therapy in patients with acute large-arteryatherosclerosis stroke.. Journal of stroke and cerebrovascular diseases : the official journal of National Stroke Association2014;23(7):1975-81- Pubmed Journal

[50] Sandercock PAG, Counsell C, Tseng M-C, Cecconi E Oral antiplatelet therapy for acute ischaemic stroke.. The Cochrane database ofsystematic reviews 2014;3 CD000029- Pubmed Journal

[51] Pan Y, Wang A, Liu G, Zhao X, Meng X, Zhao K, Liu L, Wang C, Johnston SC, Wang Y, Wang Y Cost-effectiveness of clopidogrel-aspirinversus aspirin alone for acute transient ischemic attack and minor stroke.. Journal of the American Heart Association 2014;3(3):e000912-Pubmed Journal

[52] Wong KSL, Wang Y, Leng X, Mao C, Tang J, Bath PMW, Markus HS, Gorelick PB, Liu L, Lin W, Wang Y Early dual versus mono antiplatelettherapy for acute non-cardioembolic ischemic stroke or transient ischemic attack: an updated systematic review and meta-analysis..Circulation 2013;128(15):1656-66- Pubmed Journal

[53] Whiteley WN, Adams HP, Bath PMW, Berge E, Sandset PM, Dennis M, Murray GD, Wong K-SL, Sandercock PAG Targeted use of heparin,heparinoids, or low-molecular-weight heparin to improve outcome after acute ischaemic stroke: an individual patient data meta-analysis ofrandomised controlled trials.. The Lancet. Neurology 2013;12(6):539-45- Pubmed Journal

[54] Zinkstok SM, Roos YB Early administration of aspirin in patients treated with alteplase for acute ischaemic stroke: a randomisedcontrolled trial.. Lancet (London, England) 2012;380(9843):731-7- Pubmed Journal

[55] Mihalopoulos C, Cadilhac DA, Moodie ML, Dewey HM, Thrift AG, Donnan GA, Carter RC Development and application of Model ofResource Utilization, Costs, and Outcomes for Stroke (MORUCOS): an Australian economic model for stroke.. International journal oftechnology assessment in health care 2005;21(4):499-505-null Pubmed


152 of 157


http://dx.doi.org/10.1136/neurintsurg-2014-011282


http://dx.doi.org/10.1212/WNL.0000000000002439




http://dx.doi.org/10.1080/13696998.2016.1174868

http://dx.doi.org/10.1016/S0140-6736(16)30468-8

http://dx.doi.org/10.1016/S0140-6736(16)30468-8










http://dx.doi.org/10.1161/JAHA.114.000912


http://dx.doi.org/10.1161/CIRCULATIONAHA.113.003187


http://dx.doi.org/10.1016/S1474-4422(13)70079-6


http://dx.doi.org/10.1016/S0140-6736(12)60949-0


[56] Lee M, Ovbiagele B, Hong K-S, Wu Y-L, Lee J-E, Rao NM, Feng W, Saver JL Effect of Blood Pressure Lowering in Early Ischemic Stroke:Meta-Analysis . Stroke 2015/05/28 ; Website

[57] Bath PM, Krishnan K Interventions for deliberately altering blood pressure in acute stroke. Cochrane Database of Systematic Reviews2014; Journal

[58] Qureshi AI, Palesch YY, Barsan WG, Hanley DF, Hsu CY, Martin RL, Moy CS, Silbergleit R, Steiner T, Suarez JI, Toyoda K, Wang Y,Yamamoto H, Yoon B-W Intensive Blood-Pressure Lowering in Patients with Acute Cerebral Hemorrhage.. The New England journal ofmedicine 2016; Pubmed

[59] Back L, Nagaraja V, Kapur A, Eslick GD Role of decompressive hemicraniectomy in extensive middle cerebral artery strokes: a meta-analysis of randomised trials.. Internal medicine journal 2015;45(7):711-7- Pubmed Journal

[60] Jüttler E, Unterberg A, Woitzik J, Bösel J, Amiri H, Sakowitz OW, Gondan M, Schiller P, Limprecht R, Luntz S, Schneider H, Pinzer T,Hobohm C, Meixensberger J, Hacke W Hemicraniectomy in older patients with extensive middle-cerebral-artery stroke.. The New Englandjournal of medicine 2014;370(12):1091-100- Pubmed Journal

[61] Cruz-Flores S, Berge E, Whittle IR Surgical decompression for cerebral oedema in acute ischaemic stroke. Cochrane Database ofSystematic Reviews 2012; Pubmed Journal

[62] Sandercock PA, Soane T Corticosteroids for acute ischaemic stroke. Cochrane Database of Systematic Reviews 2011; Pubmed Journal

[63] Hofmeijer J, van der Worp HB, Kappelle LJ, Eshuis S, Algra A, Greving JP Cost-effectiveness of surgical decompression for space-occupying hemispheric infarction.. Stroke; a journal of cerebral circulation 2013;44(10):2923-5- Pubmed Journal

[64] Yang J, Liu M, Zhou J, Zhang S, Lin S, Zhao H Edaravone for acute intracerebral haemorrhage. Cochrane Database of Systematic Reviews2011; Pubmed Journal

[65] Yang J, Liu M, Zhou J, Zhang S, Lin S, Zhao H Edaravone for acute intracerebral haemorrhage.. The Cochrane database of systematicreviews 2011;(2):CD007755- Pubmed Journal

[66] Bajenaru O, Tiu C, Moessler H, Antochi F, Muresanu D, Popescu BO, Novak P Efficacy and safety of Cerebrolysin in patients withhemorrhagic stroke.. Journal of medicine and life 3(2):137-43- Pubmed

[67] Sprigg N, Renton CJ, Dineen RA, Kwong Y, Bath PMW Tranexamic acid for spontaneous intracerebral hemorrhage: a randomizedcontrolled pilot trial (ISRCTN50867461).. Journal of stroke and cerebrovascular diseases : the official journal of National Stroke Association2014;23(6):1312-8- Pubmed Journal

[68] Steiner T, Poli S, Griebe M, Hüsing J, Hajda J, Freiberger A, Bendszus M, Bösel J, Christensen H, Dohmen C, Hennerici M, Kollmer J,Stetefeld H, Wartenberg KE, Weimar C, Hacke W, Veltkamp R Fresh frozen plasma versus prothrombin complex concentrate in patients withintracranial haemorrhage related to vitamin K antagonists (INCH): a randomised trial. The Lancet Neurology 15(6):566-573 Journal Website

[69] Baharoglu MI, Cordonnier C, Salman RA-S, de Gans K, Koopman MM, Brand A, Majoie CB, Beenen LF, Marquering HA, Vermeulen M,Nederkoorn PJ, de Haan RJ, Roos YB Platelet transfusion versus standard care after acute stroke due to spontaneous cerebral haemorrhageassociated with antiplatelet therapy (PATCH): a randomised, open-label, phase 3 trial.. Lancet (London, England) 2016; Pubmed Journal

[70] Mendelow AD, Gregson BA, Rowan EN, Murray GD, Gholkar A, Mitchell PM Early surgery versus initial conservative treatment inpatients with spontaneous supratentorial lobar intracerebral haematomas (STICH II): a randomised trial.. Lancet (London, England)2013;382(9890):397-408- Pubmed Journal


153 of 157





http://dx.doi.org/10.1111/imj.12724
















http://dx.doi.org/10.1016/S1474-4422(16)00110-1

http://dx.doi.org/10.1016/S1474-4422(16)00110-1


http://dx.doi.org/10.1016/S0140-6736(16)30392-0


http://dx.doi.org/10.1016/S0140-6736(13)60986-1

[71] Xiao BO, Wu F-F, Zhang H, Ma Y-B A randomized study of urgent computed tomography-based hematoma puncture and aspiration in theemergency department and subsequent evacuation using craniectomy versus craniectomy only.. Journal of neurosurgery2012;117(3):566-73- Pubmed Journal

[72] King NKK, Lai JL, Tan LB, Lee KK, Pang BC, Ng I, Wang E A randomized, placebo-controlled pilot study of patients with spontaneousintraventricular haemorrhage treated with intraventricular thrombolysis.. Journal of clinical neuroscience : official journal of theNeurosurgical Society of Australasia 2012;19(7):961-4- Pubmed Journal

[73] Gregson BA, Broderick JP, Auer LM, Batjer H, Chen X-C, Juvela S, Morgenstern LB, Pantazis GC, Teernstra OPM, Wang W-Z, ZuccarelloM, Mendelow AD Individual patient data subgroup meta-analysis of surgery for spontaneous supratentorial intracerebral hemorrhage..Stroke; a journal of cerebral circulation 2012;43(6):1496-504- Pubmed Journal

[74] Naff N, Williams MA, Keyl PM, Tuhrim S, Bullock MR, Mayer SA, Coplin W, Narayan R, Haines S, Cruz-Flores S, Zuccarello M, Brock D,Awad I, Ziai WC, Marmarou A, Rhoney D, McBee N, Lane K, Hanley DF Low-dose recombinant tissue-type plasminogen activator enhancesclot resolution in brain hemorrhage: the intraventricular hemorrhage thrombolysis trial.. Stroke; a journal of cerebral circulation2011;42(11):3009-16- Pubmed Journal

[75] Chen C-C, Liu C-L, Tung Y-N, Lee H-C, Chuang H-C, Lin S-Z, Cho D-Y Endoscopic surgery for intraventricular hemorrhage (IVH) caused bythalamic hemorrhage: comparisons of endoscopic surgery and external ventricular drainage (EVD) surgery.. World neurosurgery2011;75(2):264-8- Pubmed Journal

[76] Bennett MH, Weibel S, Wasiak J, Schnabel A, French C, Kranke P Hyperbaric oxygen therapy for acute ischaemic stroke.. The Cochranedatabase of systematic reviews 2014;11 CD004954- Pubmed Journal

[77] Ali K, Warusevitane A, Lally F, Sim J, Sills S, Pountain S, Nevatte T, Allen M, Roffe C The stroke oxygen pilot study: a randomized controlledtrial of the effects of routine oxygen supplementation early after acute stroke--effect on key outcomes at six months.. PloS one2014;8(6):e59274- Pubmed Journal

[78] Roffe C, Ali K, Warusevitane A, Sills S, Pountain S, Allen M, Hodsoll J, Lally F, Jones P, Crome P The SOS pilot study: a RCT of routineoxygen supplementation early after acute stroke--effect on recovery of neurological function at one week.. PloS one 2011;6(5):e19113-Pubmed Journal

[79] Rønning OM, Guldvog B Should stroke victims routinely receive supplemental oxygen? A quasi-randomized controlled trial.. Stroke; ajournal of cerebral circulation 1999;30(10):2033-7-null Pubmed

[80] Dávalos A, Alvarez-Sabín J, Castillo J, Díez-Tejedor E, Ferro J, Martínez-Vila E, Serena J, Segura T, Cruz VT, Masjuan J, Cobo E, Secades JJ,Citicoline in the treatment of acute ischaemic stroke: an international, randomised, multicentre, placebo-controlled study (ICTUS trial)..Lancet (London, England) 2012;380(9839):349-57-null Pubmed Journal

[81] Chamorro A, Amaro S, Castellanos M, Segura T, Arenillas J, Martí-Fábregas J, Gállego J, Krupinski J, Gomis M, Cánovas D, Carné X,Deulofeu R, Román LS, Oleaga L, Torres F, Planas AM, Safety and efficacy of uric acid in patients with acute stroke (URICO-ICTUS): arandomised, double-blind phase 2b/3 trial.. The Lancet. Neurology 2014;13(5):453-60-null Pubmed Journal

[82] Ginsberg MD, Palesch YY, Hill MD, Martin RH, Moy CS, Barsan WG, Waldman BD, Tamariz D, Ryckborst KJ, High-dose albumin treatmentfor acute ischaemic stroke (ALIAS) Part 2: a randomised, double-blind, phase 3, placebo-controlled trial.. The Lancet. Neurology2013;12(11):1049-58-null Pubmed Journal

[83] Saver JL, Kidwell C, Eckstein M, Starkman S, Prehospital neuroprotective therapy for acute stroke: results of the Field Administration ofStroke Therapy-Magnesium (FAST-MAG) pilot trial.. Stroke; a journal of cerebral circulation 2004;35(5):e106-8-null Pubmed


154 of 157


http://dx.doi.org/10.3171/2012.5.JNS111611








http://dx.doi.org/10.1016/j.wneu.2010.07.041









http://dx.doi.org/10.1016/S0140-6736(12)60813-7


http://dx.doi.org/10.1016/S1474-4422(14)70054-7


http://dx.doi.org/10.1016/S1474-4422(13)70223-0


[84] Heiss W-D, Brainin M, Bornstein NM, Tuomilehto J, Hong Z, Cerebrolysin in patients with acute ischemic stroke in Asia: results of adouble-blind, placebo-controlled randomized trial.. Stroke; a journal of cerebral circulation 2012;43(3):630-6-null Pubmed Journal

[85] Bellolio MF, Gilmore RM, Ganti L Insulin for glycaemic control in acute ischaemic stroke.. The Cochrane database of systematic reviews2014;1 CD005346- Pubmed Journal

[86] Ntaios G, Papavasileiou V, Bargiota A, Makaritsis K, Michel P Intravenous insulin treatment in acute stroke: a systematic review andmeta-analysis of randomized controlled trials.. International journal of stroke : official journal of the International Stroke Society2014;9(4):489-93- Pubmed Journal

[87] #LINN DELETE# Insulin for glycaemic control in acute ischaemic stroke [Data only. When citing this record quote "Cochrane Database ofSystematic Reviews 2014, Issue 1".].

[88] #LINN DELETE# Insulin for glycaemic control in acute ischaemic stroke [Data only. When citing this record quote "Cochrane Database ofSystematic Reviews 2014, Issue 1".].

[89] Middleton S, McElduff P, Ward J, Grimshaw JM, Dale S, D'Este C, Drury P, Griffiths R, Cheung NW, Quinn C, Evans M, Cadilhac D, Levi CImplementation of evidence-based treatment protocols to manage fever, hyperglycaemia, and swallowing dysfunction in acute stroke (QASC):a cluster randomised controlled trial.. Lancet (London, England) 2011;378(9804):1699-706- Pubmed Journal

[90] Capes SE, Hunt D, Malmberg K, Pathak P, Gerstein HC Stress hyperglycemia and prognosis of stroke in nondiabetic and diabetic patients:a systematic overview.. Stroke; a journal of cerebral circulation 2001;32(10):2426-32- Pubmed

[91] Gray CS, Scott JF, French JM, Alberti KGMM, O'Connell JE Prevalence and prediction of unrecognised diabetes mellitus and impairedglucose tolerance following acute stroke.. Age and ageing 2004;33(1):71-7-null Pubmed

[92] Kernan WN, Viscoli CM, Inzucchi SE, Brass LM, Bravata DM, Shulman GI, McVeety JC Prevalence of abnormal glucose tolerancefollowing a transient ischemic attack or ischemic stroke.. Archives of internal medicine 2005;165(2):227-33-null Pubmed

[93] Gray CS, Scott JF, French JM, Alberti KGMM, O'Connell JE Prevalence and prediction of unrecognised diabetes mellitus and impairedglucose tolerance following acute stroke.. Age and ageing 2004;33(1):71-7-null Pubmed

[94] Allport L, Baird T, Butcher K, Macgregor L, Prosser J, Colman P, Davis S Frequency and temporal profile of poststroke hyperglycemia usingcontinuous glucose monitoring.. Diabetes care 2006;29(8):1839-44-null Pubmed

[95] Vermeer SE, Sandee W, Algra A, Koudstaal PJ, Kappelle LJ, Dippel DWJ, Impaired glucose tolerance increases stroke risk in nondiabeticpatients with transient ischemic attack or minor ischemic stroke.. Stroke; a journal of cerebral circulation 2006;37(6):1413-7-null Pubmed

[96] Thompson HJ Evidence-base for fever interventions following stroke.. Stroke; a journal of cerebral circulation 2015;46(5):e98-e100-Pubmed Journal


[98] Drury P, Levi C, McInnes E, Hardy J, Ward J, Grimshaw JM, D' Este C, Dale S, McElduff P, Cheung NW, Quinn C, Griffiths R, Evans M,Cadilhac D, Middleton S Management of fever, hyperglycemia, and swallowing dysfunction following hospital admission for acute stroke inNew South Wales, Australia.. International journal of stroke : official journal of the International Stroke Society 2014;9(1):23-31- PubmedJournal


155 of 157








http://dx.doi.org/10.1016/S0140-6736(11)61485-2










http://dx.doi.org/10.1016/S0140-6736(11)61485-2



[99] Ovesen C, Brizzi M, Pott FC, Thorsen-Meyer HC, Karlsson T, Ersson A, Christensen H, Norrlin A, Meden P, Krieger DW, Petersson JFeasibility of endovascular and surface cooling strategies in acute stroke.. Acta neurologica Scandinavica 2013;127(6):399-405- PubmedJournal

[100] #LINN DELETE# Middleton S, McElduff P, Ward J, Grimshaw JM, Dale S, D'Este C, Drury P, Griffiths R, Cheung NW, Quinn C, Evans M,Cadilhac D, Levi C Implementation of evidence-based treatment protocols to manage fever, hyperglycaemia, and swallowing dysfunction inacute stroke (QASC): a cluster randomised controlled trial.. Lancet (London, England) 2011;378(9804):1699-706- Pubmed Journal

[101] Dale S, Levi C, Ward J, Grimshaw JM, Jammali-Blasi A, D'Este C, Griffiths R, Quinn C, Evans M, Cadilhac D, Cheung NW, Middleton SBarriers and enablers to implementing clinical treatment protocols for fever, hyperglycaemia, and swallowing dysfunction in the Quality inAcute Stroke Care (QASC) Project--a mixed methods study.. Worldviews on evidence-based nursing / Sigma Theta Tau International, HonorSociety of Nursing 2015;12(1):41-50- Pubmed Journal

[102] den Hertog HM, van der Worp HB, van Gemert HMA, Algra A, Kappelle LJ, van Gijn J, Koudstaal PJ, Dippel DWJ The Paracetamol(Acetaminophen) In Stroke (PAIS) trial: a multicentre, randomised, placebo-controlled, phase III trial.. The Lancet. Neurology2009;8(5):434-40- Pubmed Journal

[103] Den Hertog HM, van der Worp HB, Tseng M-C, Dippel DW Cooling therapy for acute stroke.. The Cochrane database of systematicreviews 2009;(1):CD001247- Pubmed Journal

[104] Castillo J, Dávalos A, Marrugat J, Noya M Timing for fever-related brain damage in acute ischemic stroke.. Stroke; a journal of cerebralcirculation 1998;29(12):2455-60-null Pubmed

[105] Greer DM, Funk SE, Reaven NL, Ouzounelli M, Uman GC Impact of fever on outcome in patients with stroke and neurologic injury: acomprehensive meta-analysis.. Stroke; a journal of cerebral circulation 2008;39(11):3029-35- Pubmed Journal

[106] Langhorne P, Stott DJ, Robertson L, MacDonald J, Jones L, McAlpine C, Dick F, Taylor GS, Murray G Medical complications after stroke: amulticenter study.. Stroke; a journal of cerebral circulation 2000;31(6):1223-9-null Pubmed

[107] Chen Y-W, Chang K-H, Chen H-C, Liang W-M, Wang Y-H, Lin Y-N The effects of surface neuromuscular electrical stimulation on post-stroke dysphagia: a systemic review and meta-analysis.. Clinical rehabilitation 2016;30(1):24-35- Pubmed Journal

[108] Huang K-L, Liu T-Y, Huang Y-C, Leong C-P, Lin W-C, Pong Y-P Functional outcome in acute stroke patients with oropharyngeal Dysphagiaafter swallowing therapy.. Journal of stroke and cerebrovascular diseases : the official journal of National Stroke Association2014;23(10):2547-53- Pubmed Journal


[110] Bath PM, Scutt P, Love JO, Clavé P, Cohen D, Dziewas R, Iversen HK, Ledl C, Ragab S, Soda H, Warusevitane A, Woisard V, Hamdy S,Pharyngeal Electrical Stimulation for Treatment of Dysphagia in Subacute Stroke: A Randomized Controlled Trial.. Stroke; a journal ofcerebral circulation 2016;47(6):1562-70-null Pubmed Journal

[111] Scutt P, Lee HS, Hamdy S, Bath PM Pharyngeal Electrical Stimulation for Treatment of Poststroke Dysphagia: Individual Patient DataMeta-Analysis of Randomised Controlled Trials.. Stroke research and treatment 2015;2015 429053-null Pubmed Journal

[112] Suntrup S, Marian T, Schröder JB, Suttrup I, Muhle P, Oelenberg S, Hamacher C, Minnerup J, Warnecke T, Dziewas R Electricalpharyngeal stimulation for dysphagia treatment in tracheotomized stroke patients: a randomized controlled trial.. Intensive care medicine2015;41(9):1629-37-null Pubmed Journal


156 of 157


http://dx.doi.org/10.1111/ane.12059


http://dx.doi.org/10.1016/S0140-6736(11)61485-2


http://dx.doi.org/10.1111/wvn.12078


http://dx.doi.org/10.1016/S1474-4422(09)70051-1








http://dx.doi.org/10.1177/0269215515571681




http://dx.doi.org/10.1016/S0140-6736(11)61485-2




http://dx.doi.org/10.1155/2015/429053


http://dx.doi.org/10.1007/s00134-015-3897-8

[113] Long Y-B, Wu X-P A meta-analysis of the efficacy of acupuncture in treating dysphagia in patients with a stroke.. Acupuncture inmedicine : journal of the British Medical Acupuncture Society 2012;30(4):291-7-null Pubmed Journal

[114] Geeganage C, Beavan J, Ellender S, Bath PMW Interventions for dysphagia and nutritional support in acute and subacute stroke.. TheCochrane database of systematic reviews 2012;10 CD000323-null Pubmed Journal

[115] Bakhtiyari J, Sarraf P, Nakhostin-Ansari N, Tafakhori A, Logemann J, Faghihzadeh S, Harirchian MH Effects of early intervention ofswallowing therapy on recovery from dysphagia following stroke.. Iranian journal of neurology 2015;14(3):119-24-null Pubmed

[116] Pisegna JM, Kaneoka A, Pearson WG, Kumar S, Langmore SE Effects of non-invasive brain stimulation on post-stroke dysphagia: Asystematic review and meta-analysis of randomized controlled trials.. Clinical neurophysiology : official journal of the International Federationof Clinical Neurophysiology 2016;127(1):956-68-null Pubmed Journal

[117] Du J, Yang F, Liu L, Hu J, Cai B, Liu W, Xu G, Liu X Repetitive transcranial magnetic stimulation for rehabilitation of poststroke dysphagia:A randomized, double-blind clinical trial.. Clinical neurophysiology : official journal of the International Federation of Clinical Neurophysiology2016;127(3):1907-13-null Pubmed Journal

[118] Crary MA, Carnaby GD, Sia I, Khanna A, Waters MF Spontaneous swallowing frequency has potential to identify dysphagia in acutestroke.. Stroke; a journal of cerebral circulation 2013;44(12):3452-7-null Pubmed Journal


157 of 157


http://dx.doi.org/10.1136/acupmed-2012-010155





http://dx.doi.org/10.1016/j.clinph.2015.04.069


http://dx.doi.org/10.1016/j.clinph.2015.11.045



Guideline Acute medical and surgical management- …...Acute medical and surgical management- Clinical Guidelines for Stroke Management (2017) Contact Language en Start Date 04.09.2016

Documents