5. Acute Ischemic Stroke Treatment

Treatment benefits from revascularization decrease over time, and 1.9 million brain cells die every minute following stroke onset (Saver, 2006); therefore, all patients with stroke should be treated as quickly as possible to maximize the potential for the best outcomes. The new extended time windows should not be interpreted to mean that time to treatment should be slowed in any way.

Acute stroke treatment is shifting from a time-based to a tissue-based paradigm as emerging evidence suggests that the speed of stroke progression differs between individuals and beneficial treatment may be offered beyond standard time windows. Using time as the sole criterion to select patients for thrombolysis and endovascular thrombectomy (EVT) may lead to missed opportunities for treatment. Nevertheless, “time is brain” remains a reality, and delays in stroke diagnosis and treatment are associated with worse outcomes” (Desai and Smith, Cardiovascular Disease and Stroke, 2013).

Endovascular intervention refers to EVT as well as other endovascular procedures such as stenting.

Recommendations and/or Clinical Considerations
5.1 Patient Selection for Acute Ischemic Stroke Treatments
  1. Within 6 hours of stroke symptom onset: All patients with disabling acute ischemic stroke who can be treated within the indicated time windows must be screened without delay by a physician with stroke expertise (either on-site or by virtual acute stroke care/telestroke consultation) to determine their eligibility for both intravenous thrombolysis and/or interventional treatment with EVT within a 6-hour window from stroke symptom onset or last known well time) [Strong recommendation; High quality of evidence].
    1. When it is unclear if a patient should be treated with IV thrombolysis, urgent consultation with a stroke specialist on-site or through virtual stroke services is recommended [Strong recommendation; Low quality of evidence]
    2. If there is uncertainty about interpretation of CT imaging, urgent consultation with a radiologist either on-site or through virtual stroke services is recommended [Strong recommendation; Low quality of evidence].

    Section 5.1 Clinical Considerations

    1. Intravenous thrombolysis beyond 4.5 hours may be considered, in consultation with a physician with stroke expertise and based on advanced imaging.
    2. If a large vessel occlusion (LVO) is present, consideration for thrombolysis beyond 4.5 hours from the time the patient was last known well should not delay decisions regarding EVT.
    5.2 (REVISED FOR 2022) Intravenous Thrombolysis Administration

    1. All eligible patients with disabling ischemic stroke, who can receive intravenous thrombolysis with either alteplase or tenecteplase within 4.5 hours of stroke symptom onset time or last known well time should be offered intravenous thrombolysis [Strong recommendation; High quality of evidence]

    1. Treatment should be initiated as soon as possible after patient arrival and CT scan completion [Strong recommendation; High quality of evidence].
    2. Every effort should be made to ensure door-to-needle times are routinely monitored and improved [Strong recommendation; Moderate quality of evidence].
    1. For patients with sudden deterioration during or following administration of IV thrombolysis, an emergent CT scan should be done [Strong recommendation; Moderate quality of evidence].
    2. For patients with orolingual angio-edema:
      1. IV thrombolysis should be discontinued if still infusing at the first signs of angioedema [Strong recommendation; Moderate quality of evidence].
      2. The following medications are recommended: antihistamines (H1 blocker [e.g., diphenhydramine], H2 blocker [e.g., famotidine]). Consider glucocorticoids inhaled racemic epinephrine as part of standard airway management [Strong recommendation; Low quality of evidence].

      For patients with symptomatic ICH following IV thrombolysis refer to section 5.6.

      1. Consultation with appropriate specialists should be undertaken to aid in achieving hemostasis [Strong recommendation; Low quality of evidence].

      Section 5.2 Clinical Considerations

      1. Consent: Intravenous thrombolysis and EVT are considered the standard of care for acute stroke treatment. Routine procedures for emergency consent apply.
      2. Intravenous thrombolytic administration for patients on DOACs: Intravenous thrombolytics should not routinely be administered to patients on DOACs who present with acute ischemic stroke. In comprehensive stroke centres with access to specialized tests of DOAC levels and reversal agents, thrombolysis could be considered, and decisions should be based on individual patient characteristics, in consultation with thrombosis specialists, patients, and their families.
        1. The benefits and risks of providing intravenous thrombolysis to a patient who is being treated with the combination of antiplatelet and low-dose DOAC (i.e., COMPASS trial protocol) are unclear. Treatment may be considered in consultation with a stroke expert.
        2. Anticoagulation is not a contraindication for EVT, and the decision to treat should be based on individual patient factors and assessment of benefit and risk.
        3. Patients who present with stroke who are taking a DOAC may be considered for rapid reversal if otherwise eligible for IV thrombolysis and if a reversal agent is readily available. Consultation with an expert in stroke care in strongly advised for these cases.
        1. For example, this may apply to pediatric patients with stroke (newborn to age 18 years); and pregnant women who experience an acute ischemic stroke. Refer to Canadian Stroke Best Practices Management of Acute Stroke During Pregnancy Consensus Statement for additional information.
        5.3 Stroke While Already in Hospital

        1. Patients already admitted to hospital* who present with a sudden onset of new stroke symptoms should be rapidly evaluated without delay for eligibility for acute stroke treatment and provided with access to appropriate acute stroke treatments (including IV thrombolysis and EVT) [Strong recommendation; Moderate quality of evidence].

        Note: When an inpatient has a stroke while in hospital, all other sections of the CSBP modules apply to these patients for assessment, diagnosis, management, and recovery.

        5.4 Endovascular Thrombectomy for Acute Ischemic Stroke

        Refer to Section 4.2 and Boxes 4A, 4B, and 4C for detailed recommendations on neuroimaging-based selection criteria. Refer to Box 5D for information on the Pre- and Post- Management of Patients Undergoing Endovascular Thrombectomy.

        1. Endovascular Thrombectomy (EVT) should be offered within a coordinated system of care including coordination among emergency medical services, access to rapid neurovascular (brain and vascular) imaging, the emergency department, the stroke team and radiology, local experts in neuro intervention, anesthesia, and access to a stroke unit for ongoing management [Strong recommendation; High quality of evidence].
        2. EVT is indicated in patients based on imaging selection, most commonly performed with non-contrast CT head and CT angiography (including extracranial and intracranial arteries) [Strong recommendation; High quality of evidence]. Refer to Box 5C for inclusion criteria for EVT.
        3. EVT may be indicated in patients with proximal anterior circulation occlusions who have received intravenous thrombolysis, as well as those who are not eligible for intravenous thrombolysis [Strong recommendation; High quality of evidence].
        4. Intravenous thrombolysis should be provided to all eligible patients, including those patients who are also eligible for EVT [Strong recommendation; High quality of evidence].
          1. For patients who are also eligible for intravenous thrombolysis, this should be initiated while simultaneously preparing the angiography suite for EVT [Strong recommendation; High quality of evidence]. Treatment with either intravenous thrombolysis or EVT should not be delayed for any reason.

          5.4.1 Anterior Circulation

          1. For large artery occlusions in the anterior circulation, EVT should be considered based on patient pre-morbid function, clinical deficit, and imaging findings. Patients who can be treated within 6 hours of symptom onset (i.e., arterial access within 6 hours of last known well time) should receive EVT [Strong recommendation; High quality of evidence]. Refer to Box 4B for Imaging Inclusion Criteria for endovascular thrombectomy.
          2. Selected patients with LVO and who are eligible based on premorbid status and advanced neuroimaging, should be treated with EVT within 24 hours of last known well time (i.e., arterial access within 24 hours of last known well time) [Strong recommendation; High quality of evidence]. Refer to Box 4C for imaging inclusion criteria for EVT beyond 6 hours from stroke symptom onset.

          5.4.2 Posterior Circulation

          1. For large artery occlusions in the posterior circulation (e.g., basilar artery occlusion) EVT should be considered based on patient pre-morbid function, clinical deficit, and imaging findings. Consultation with a physician with stroke expertise and with the patient and/or substitute decision-makers is recommended [ Conditional recommendation; Moderate quality of evidence ]. Note: Randomized trials are ongoing, and this guidance will be reviewed when trial results are available.

          5.4.3 Sedation for Endovascular Interventions

          1. For endovascular interventions, procedural sedation is generally preferred over intubation and general anesthesia in most patients [Strong recommendation; Moderate quality of evidence]
          2. General anesthesia is appropriate if medically indicated (e.g., for airway compromise, respiratory distress, depressed level of consciousness, severe agitation, or other indication potentially impairing the technical ability to perform the procedure, as determined by the treating physician). General anesthesia may also be considered if technical complexity is expected during the stroke intervention. In such cases, excessive and prolonged hypotension and time delays should be avoided [Strong recommendation; Moderate quality of evidence].

          Section 5.4 Clinical Considerations

          1. For patients transferred to an EVT-enabled hospital, repeat neuroimaging immediately on arrival, to confirm eligibility, may be considered. The decision to repeat may be based on multiple factors: initial imaging features (including quality), clinical presentation, adjuvant medical therapies, changes in health status, and delay in arrival to the EVT-enabled site. Repeat imaging may include part or all of the neuroimaging recommended in section 4.2.
          2. Device selection should be at the discretion of the interventionalists based on clinical and technical factors during the procedure.
          3. There should be a process at EVT centres to activate anesthesia without delay when deemed necessary.
          4. Patients with stroke discovered on awakening or with unknown last known well time should be considered for EVT if eligible based on imaging findings and clinical presentation. Refer to Box 5C for more information.
          5. For patients undergoing EVT following administration of thrombolysis, there should not be a delay in proceeding to EVT to determine clinical effectiveness of thrombolysis.
          6. (NEW FOR 2022) When a patient who is eligible for both intravenous thrombolysis and EVT presents DIRECTLY TO AN EVT-CAPABLE HOSPITAL, a decision not to administer intravenous thrombolysis and proceed straight to EVT must balance both the patient-related and operational factors in play at that moment, for that patient. The overarching focus is to improve patient outcomes while safely reducing door-to-needle and door-to-puncture times. The main driver for an excellent outcome remains “time is brain.”

          Note: Clinical consideration 6 is controversial. It will be updated as additional evidence becomes available. In the meantime, clinicians involved in acute stroke care should focus on improving patient outcomes while safely reducing door-to-needle and door-to-puncture times. The main driver for excellent outcomes remains “time is brain.”.

          5.5 Seizure Management
          1. Seizure in the presence of suspected acute stroke is not a contraindication for revascularization and could be treated using appropriate short-acting medications (e.g., lorazepam IV) if the seizures are not self-limited [ Conditional recommendation ; Low quality of evidence].
          5.6 (NEW FOR 2022) Emergency Management of Thrombolysis-Associated Hemorrhage

          Note: Section 5.6 applies to patients experiencing a cerebral or systemic hemorrhage following administration of intravenous thrombolysis. Refer to CSBPR guidelines on Management of Intracerebral Hemorrhage for additional information.

          5.6.1 Intracranial Hemorrhage

          1. Intracranial hemorrhage should be considered if there is a change in neurological symptoms or signs, especially a reduction in level of consciousness, or a spike in blood pressure with persisting blood pressure elevation, or new or worsened headache [Strong recommendation; Moderate quality of evidence].
          2. An immediate non-contrast CT head should be done to assess for intracranial hemorrhage [Strong recommendation; Moderate quality of evidence].
          3. The patient should be accompanied to the CT by a member of the stroke team and the results reviewed immediately. If there is no intracranial hemorrhage, CTA should be urgently considered to identify intracranial occlusion and the need for urgent EVT should be considered [Strong recommendation; Moderate quality of evidence].
          4. If intracranial hemorrhage is identified, the intravenous thrombolysis infusion should be discontinued immediately if it is still running [Strong recommendation; Moderate quality of evidence].
          5. If intracranial hemorrhage is identified, blood work, including a complete blood count (CBC) and INR (PT), as well as type and cross, should be drawn [Strong recommendation; Moderate quality of evidence], with STAT results requested [Strong recommendation; Low quality of evidence].
          6. The following agents may be considered, as they have shown potential benefit and limited harm: cryoprecipitate, human fibrinogen concentrate fresh frozen plasma, tranexamic acid. Use of these medications should be considered on an individual, case-by-case basis [ Conditional recommendation ; Low quality of evidence].
          7. The following treatment options should probably be avoided as they have notshown benefit and have shown potential for harm: prothrombin complex concentrates, platelet transfusions, factor VIIa [ Conditional recommendation ; Low quality of evidence].

          5.6.2 Extracranial (Systemic) Hemorrhage Management

          Note: For systemic hemorrhage, follow local protocol for management guidance.

          1. A diagnosis of systemic bleeding should be considered when the following are present or suspected [Strong recommendation; Moderate quality of evidence]:
            1. Visible bleeding at a compressible site
            2. Reduction in blood pressure, localized pain, diaphoresis, or other signs of hypovolemic shock.

            Section 5.6 Clinical Considerations

            1. Hypertension with symptomatic ICH: In patients with secondary ICH who are hypertensive (>185/110 mm HG), blood pressure should be lowered, however, the specific target and duration of therapy are unknown at this time.

            Meta-analyses of randomized controlled trials (RCTs) of intravenous alteplase for acute ischemic stroke have shown the treatment can reduce the risk of disability and death, despite a small risk of serious bleeding. The widest time window for alteplase administration after stroke onset remains imprecisely defined, but currently available data show clear evidence of benefit when given up to 4.5 hours after the onset of symptoms. There remains a strong inverse relationship between treatment delay and clinical outcome; therefore, eligible patients should be treated as soon as possible.

            Endovascular treatment for large artery ischemic stroke in the anterior circulation has clearly demonstrated efficacy with numbers needed to treat (NNT) of approximately four to achieve functional independence at 90 days, although data from the DAWN trial (Nogueira et al. New Eng J Med 2018; 378:11-21.) suggest the NNT may be as low as three. Pooled results from a series of older trials, indicated the NNT for this outcome was higher, closer to five (HERMES, Goyal et al. Lancet 2016;387(10029):1723-31). This type of therapy can have a profoundly positive impact on patients who suffer the most devastating ischemic strokes, who without treatment may be left permanently disabled.

            People with lived experience expressed the importance of having conversations with loved ones before health changes occur. Without these conversations, it can be difficult to know what the person would want. Some individuals shared that those conversations, even around desire to be involved in trials and/or experimental treatments, can be helpful to have in advance of changes in health status. Similarly, people with lived experience stated the importance of sharing personal values and preferences with the healthcare team, to help with treatment decisions.

            (Note: DAWN trial criteria: To obtain mRS of 0-2 at 90 days (49% vs. 13%=NNT of 2.8); HERMES 2016 meta-analysis to obtain mRS score of 0-2 at 90 days (46% vs. 26.5%=NNT of 5.1).

            To ensure people experiencing a stroke receive timely stroke assessments, interventions and management, interdisciplinary teams need to have the infrastructure and resources required. These may include the following components established at a systems level.

            1. Local protocols should prioritize patients with stroke for immediate access to appropriate diagnostics such as CT imaging and neurovascular imaging with CTA. This should include patients with known times of stroke symptom onset (or time last known well), and patients who are discovered with stroke symptoms on wakening.
            2. Coordinated and integrated systems of care involving all relevant personnel in the prehospital and emergency care of patients with stroke, including paramedics, emergency department staff, stroke teams, radiologists and neurointerventionists. Protocols should be in place in partnership with EMS agencies and treating hospitals, and between hospitals within stroke systems to ensure rapid transport to centres providing advanced stroke services within treatment time windows.
            3. Consideration should be given to northern, rural, remote, and Indigenous residents to ensure immediate access to appropriate diagnostics and treatment is not delayed.
            4. (NEW FOR 2022) Health system leaders should work with hospitals to coordinate plans for launching tenecteplase and consider group purchase opportunities. Health Canada should undertake rapid review and approval for use of tenecteplase in acute stroke.
            5. Health regions and stroke systems should examine and determine the possible resource impact of the EVT time window extension (up to 24 hours in highly selected cases). Demand for imaging will increase especially at comprehensive stroke and EVT-enabled centres. Staffing, service hours and capacity should be considered to ensure efficiency and effectiveness of services.
            6. System planners and patient flow specialists should plan for significant challenges associated with diversion of potential EVT candidates to EVT-enabled centres. This will affect emergency departments, radiology departments, and acute inpatient units where occupancy rates are often stretched (over 100% in many hospitals).
            7. Stroke neurology and neurointerventional expertise should be regionalized, with a system in place across regions for rapid access to physicians experienced in acute thrombolysis and endovascular therapies, including through telemedicine. This includes protocols for contacting physicians with stroke expertise for administration of intravenous thrombolysis, as well as transport to higher levels of stroke care, as needed, for emergent treatments.
            8. Build capacity for trained neurointerventionists within health regions and academic institutions to ensure sufficient availability to meet regional and provincial EVT healthcare needs.
            9. Urgent acute protocols in place and well-communicated to all healthcare practitioners within the hospital regarding management of in-hospital patients with stroke, ensuring access to CT imaging of the brain, and CTA of the extracranial and intracranial vessels as soon as possible after stroke symptom onset.
            10. Access to specialized acute stroke units where staff are experienced in managing patients who have received intravenous thrombolysis or EVT.
            11. Endovascular interventional programs are in evolution across Canada; decisions around appropriate site, transfer and bypass protocols, and timelines will be determined at the provincial or regional level. Decisions about when those services are fully operational, and who should be transferred by paramedics to those facilities should be made at the provincial or regional level and communicated to all relevant stakeholders.
            12. Availability of helical CT scanners with appropriate programming for CTA (multiphase or dynamic CTA) and CTP sequences, and appropriate post-processing software optimized for the production of high-quality imaging.
            13. Monitor access, outcomes, and key processes of care that support timeliness of intervention.
            14. Ensuring ongoing capacity of EVT-capable hospitals as EVT volumes grow (i.e., ensure appropriate accountabilities for all providers accessing or providing services, ensure funding keeps pace, and ensure ability to ensure safe and timely care is maintained).

            System Indicators:

            1. Proportion of patients in rural or remote communities who receive intravenous thrombolysis through the use of telestroke technology, as a proportion of all patients with ischemic stroke in that community and as a proportion of all telestroke consults for ischemic stroke.

            Process indicators:

            1. Overall proportion of all patients with ischemic stroke who receive treatment with intravenous thrombolysis (core).
            2. Median time (in minutes) from patient arrival in the emergency department to administration of intravenous thrombolysis.
            3. Proportion of patients with ischemic stroke who receive treatment with intravenous thrombolysis within 3.0 and 4.5 hours of symptom onset.
            4. Proportion of all thrombolyzed patients with stroke who receive thrombolysis within 30 minutes of hospital arrival (core).
            5. Proportion of all patients with ischemic stroke who receive treatment with EVT (core).
            6. Median time from hospital arrival to arterial puncture, and from CT scan (first slice of the non-contrast CT) to arterial puncture for patients undergoing EVT.
            7. Median time from hospital arrival to first reperfusion for patients undergoing EVT. Time of first reperfusion is defined as the first angiographic image showing partial or complete reperfusion of the affected arterial territory. (CIHI Stroke Special Project for EVT440 indicator. Refer to Measurement Note ii below.)
            8. Proportion of patients with stroke who receive BOTH intravenous thrombolysis and EVT.
            9. For patients with stroke that occurs while in hospital for other medical reasons (in-hospital strokes), median time from last known well to brain imaging.
            10. For patients with stroke while in hospital for other medical reasons (in-hospital strokes), median time from last known well to acute thrombolysis or EVT (arterial puncture).
            11. Final reperfusion status for patients undergoing endovascular reperfusion therapy, quantified using the modified Thrombolysis in Cerebral Infarction (mTICI) system. (CIHI Stroke Special Project for EVT440 indicator. Refer to Measurement Note ii below.)
            12. Proportion of patients with symptomatic subarachnoid or intracerebral hemorrhage following intravenous thrombolysis or EVT (defined as any PH1, PH2, RIH, SAH, or IVH associated with a four-point or more worsening on the NIHSS within 24 hours).
            13. Virtual stroke care: Proportion of people with acute stroke who receive a virtual care consult with a stroke expert at another site who receive acute thrombolysis as a result.
            14. Virtual stroke care: Proportion of stroke patients managed with Telestroke who received tPA, who had a symptomatic secondary intracerebral hemorrhage, systemic hemorrhage, died in hospital, and were discharged to long-term care vs. home or to rehabilitation.

            Patient-oriented outcome and experience indicators:

            1. Level of functional impairment following acute stroke - ability to perform activities of daily living independently.
            2. Modified Rankin Scale (mRS) score of all patients with stroke who receive intravenous thrombolysis or EVT at 30 days and at 90 days following hospital discharge.
            3. Proportion of patients with symptomatic subarachnoid or intracerebral hemorrhage following EVT (defined as any PH1, PH2, RIH, SAH, or IVH associated with a four-point or more worsening on the NIHSS within 24 hours).
            4. In-hospital mortality rates (overall and 30-day) for patients with ischemic stroke, stratified by those who receive intravenous thrombolysis or EVT and those who do not.

            Measurement Notes

            1. Refer to the Quality of Stroke Care in Canada Key Quality Indicators and Stroke Case Definitions 7th Edition for calculations, process timelines, and outcome measures for intravenous thrombolysis and EVT.
            2. The denominator for treatment indicators is the number of people discharged from the emergency department (CIHI NACRS) or inpatient care (CIHI DAD) with a diagnosis of ischemic stroke – see data dictionary for relevant ICD10 codes.
            3. The Canadian Institute of Health Information has a stroke quality of care special project (#440) as part of the Discharge Abstract Database extraction that enables data collection on six performance measures for EVT. Referred to in performance measures 7 and 10 above as CIHI Stroke Special Project for EVT440.
            4. Data may be obtained from patient charts, through chart audit, registry or review.
            5. Time interval measurements should be taken from the time the patient is triaged at the hospital (according to emergency department standards of care, triage should come before registration. Triage time should always be used to standardize) until the time of intravenous thrombolysis administration noted in the patient chart (nursing notes, emergency department record, or medication record).
            6. For performance measures related to time intervals: Calculate all percentiles and examine 50th and 90th percentiles and inter-quartile range.
            7. When recording if intravenous thrombolysis is given, include times for both the administration of the bolus (both alteplase and TNK), and the time when the infusion is started (alteplase), as there are often delays between bolus and infusion which may decrease alteplase efficacy. The route of administration should also be recorded, as there are different times to administration benchmarks for intravenous and endovascular routes.
            8. For EVT, treatment time should be time of first arterial puncture.

            Resources and tools listed below that are external to Heart & Stroke and the Canadian Stroke Best Practice Recommendations may be useful resources for stroke care. However, their inclusion is not an actual or implied endorsement by the Canadian Stroke Best Practices writing group. The reader is encouraged to review these resources and tools critically and implement them into practice at their discretion.

            Healthcare Provider Information

            • Canadian Stroke Best Practice Recommendations Acute Stroke Management Module Appendix 3, Table 4: Screening and Assessment Tools for Acute Stroke Severity
            • Canadian Stroke Best Practices Management of Acute Stroke during Pregnancy Consensus Statement
            • Refer to Box 5A for Criteria for Stroke Centres Providing Acute Ischemic Stroke Treatment
            • Refer to Box 5B for Inclusion and exclusion criteria for intravenous thrombolysis eligibility.
            • Refer to Box 5C for Inclusion Criteria for endovascular therapy
            • Heart & Stroke: Virtual Stroke Care Implementation Toolkit
            • Heart & Stroke: Taking Action for Optimal Community and Long-Term Stroke Care (TACLS) A Resource for Healthcare Providers
            • American College of Chest Physicians (ACCP) Anticoagulation Guidelines
            • ASPECTS
            • Stroke Engine
            • Heart & Stroke: Acute Ischemic Stroke Intravenous Thrombolysis Protocol Checklist
            • Heart & Stroke: Acute Ischemic Stroke Endovascular Thrombectomy Protocol Checklist
            • Heart & Stroke: 2022 Acute Stroke Management Key Quality Indicators
            • The Ontario Acute Stroke Best Practice Coordinators and the Ontario Regional Education Group in collaboration with Heart & Stroke: Key Changes to the 2022 Acute Stroke Management Canadian Stroke Best Practice Recommendations

            Information for people with lived experience of stroke, including family, friends and caregivers

            • Heart & Stroke: Your Stroke Journey
            • Heart & Stroke: Post-Stroke Checklist
            • Heart & Stroke: Acute Stroke Management Infographic
            • Heart & Stroke: Taking Action for Optimal Community and Long-Term Stroke Care (TACLS) A Resource for Healthcare Providers
            • Stroke Engine

            Evidence Table and Reference List 5 (Acute Ischemic Stroke Treatment – Thrombolytic Therapy)

            Evidence Table and Reference List 5b (Acute Ischemic Stroke Treatment – Endovascular Therapy)

            The weight of evidence from many large, international trials over a time frame of 20 years clearly indicate that treatment with intravenous alteplase reduces the risk of death or disability following ischemic stroke, at 3 to 6 months post-treatment. The NINDS trial (1995) was one of the earliest, large trials conducted in the USA. Patients were randomized to receive alteplase or placebo within three hours of symptom onset. At 3 months, significantly more patients in the t-PA group had experienced a good outcome (using any one of the study’s four metrics), with no difference in 90-day mortality between groups. In contrast, patients who received alteplase within 3 to 5 hours in the ATLANTIS trial (1999) were no more likely to have a good neurological or functional outcome at 90 days than patients in the placebo group.

            The Third International Stroke Trial (IST-3, 2012) is the largest (n=3,035) and most recent trial of alteplase, in which patients were randomized to receive a standard dose of alteplase (0.9 mg/kg) or placebo. Investigators aimed to assess the risks and benefits of treatment among a broader group of patients, and determine if particular subgroups of patients might benefit preferentially from treatment. In this trial, 95% of patients did not meet the strict licensing criteria, due to advanced age or time to treatment. Unlike all previous large trials, which excluded them, IST-3 included patients >80 years of age. In fact, the majority of patients (53%) were >80 years of age. Approximately one-third of all patients were treated within 0-3 hours, 3.0-4.5 hours, and 4.5-6.0 hours of onset of symptoms. Overall, there was an increase in the risk of death within 7 days in patients who had received alteplase, although there was no difference in 6-month mortality in both crude and adjusted analyses. There was no significant difference in the percentage of patients who were treated with alteplase who were alive and independent (defined as an Oxford Handicap Score of 0-1) at 6 months (37% vs. 35%, adjusted OR=1.13, 95% CI 0.95 to 1.35, p=0.181, although a secondary ordinal analysis suggested a significant, favourable shift in the distribution of OHS scores at 6 months. Significantly improved odds of a good outcome at 6 months were associated with the subgroups of older patients (≥80 years), higher NIHSS scores, higher baseline probability of good outcome and treatment within 3 hours. Fatal or non-fatal symptomatic intracranial hemorrhage within 7 days occurred more frequently in patients in the t-PA group (7% vs. 1%, adjusted OR=6.94, 95% CI 4.07 to 11.8, p<0.0001). The 3-year risk of mortality (2016) was similar between groups (47% vs. 47%, 95% CI 3.6%, 95% CI -0.8 to 8.1); however, patients who received rt-PA had a significantly lower risk of death between 8 days and 3 years (41% vs. 47%; HR= 0.78, 95% CI 0·68–0·90, p=0·007).

            Emberson et al. (2014) used data from 6,756 patients from 9 major t-PA trials (NINDs a/b, ECASS I/II, III, ATLANTIS a/b, EPITHET, IST-3) to examine the effect of timing of administration more closely. Earlier treatment was associated with the increased odds of a good outcome, defined as an (mRS score of 0-1 (≤3.0 h: OR=1.75, 95% CI 1.35-2.27 vs. >3 to ≤4.5 h: OR=1.26, 95% CI 1.05-1051 vs. >4.5 h: OR=1.15, 95% CI 0.95-1.40). Framed slightly differently, when patient-level data from the same 9 major randomized controlled trials (RCTs) were recently pooled, Lees et al. (2016) reported that for each patient treated within 3 hours, significantly more would have a better outcome (122/1,000, 95% CI 16-171), whereas for each patient treated >4.5 hours, only 20/1,000 (95% CI -31-75, p=0.45) would have a better outcome. Wardlaw et al. (2013), including the results from 12 RCTs (7,012 patients), concluded that for every 1,000 patients treated up to 6 hours following stroke, 42 more patients were alive and independent (mRS<2) at the end of follow-up, despite an increase in early ICH and mortality. The authors also suggested that patients who did not meet strict licensing criteria due to age and timing of treatment (i.e., patients from the IST-3) trial were just as likely to benefit; however, early treatment, within 3 hours of stroke onset, was more effective.

            Results from several recent trials indicate that thrombolysis with t-PA can be used for patients outside of the previously established therapeutic window. In the Extending the Time for Thrombolysis in Emergency Neurological Deficits (EXTEND) trial (Ma et al., 2019), 225 patients with an ischemic stroke were included, where symptom onset was estimated to be 4.5 to ≤9 hours previously. Recruitment was suspended after the results of the WAKE-UP trial became available. The primary outcome (mRS 0-1 at 90 days) occurred in 35.4% of the patients in the alteplase group and 29.5% in the control (placebo) group. After adjustment for age and baseline severity, the likelihood of the primary outcome significantly increased in the alteplase group (RR=1.44, 95% CI 1.01–2.06), as did the proportion of patients who attained a mRS score of 0-2 at 90 days (49.6% vs. 42.9%; adjusted RR=1.36, 95% CI, 1.06 to 1.76); however there was no significant difference between groups in functional improvement at 90 days (i.e., shift in mRS scores; RR=1.55, 95% CI 0.96 to 2.49). The results from the Efficacy and Safety of MRI-based Thrombolysis in Wake-up Stroke (WAKE-Up) trial (Thomalla et al., 2018) also suggest that highly selected patients with mild to moderate ischemic strokes and an unknown time of symptom onset, treated with alteplase may also benefit from treatment. Patients in this trial were not eligible for treatment with mechanical thrombectomy and were selected based on a pattern of DWI-FLAIR-mismatch. A significantly higher proportion of patients in the alteplase group had a favourable clinical outcome (mRS 0-1) at 90 days (53.3% vs. 41.8%, adj OR=1.61, 95% CI 1.06-2.36, p=0.02), although the risk of type 2 parenchymal hemorrhage was significantly higher compared with placebo (4% vs. 0.4%, adj OR=10.46, 95% CI 1.32 to 82.77, p=0.03).

            The standard treatment dose of rt-PA is established to be 0.9 mg/kg, with a maximum dose of 90 mg. The non-inferiority of a lower dose (0.6 mg/kg) was recently examined in the Enhanced Control of Hypertension and Thrombolysis Stroke Study (ENCHANTED) trial (Anderson et al., 2016). The primary outcome (death or disability at 90 days) occurred in 53.2% of low-dose patients and 51.1% in standard-dose patients (OR=1.09, 95% CI 0.95-1.25, p for non-inferiority=0.51), which exceeded the upper boundary set for non-inferiority of 1.14. The risks of death within 90 days or serious adverse events did not differ significantly between groups (low dose vs. standard dose: 8.5% vs. 10.3%; OR=0.80, 95% CI 0.63-1.01, p=0.07 and 25.1% vs. 27.3%; OR=0.89, 95% CI 0.76-1.04, p=0.16, respectively), although the risk of symptomatic ICH was significantly higher in patients that received the standard dose of rt-PA.

            Strategies to improve guideline adherence have been shown to help improve thrombolysis uptake and shorten thrombolysis process times. In Canada, following the initiation of an Improvement Collaborative intervention during 2016–2017, the number of patients receiving thrombolysis increased from 9.35% in the pre-period to 15.73% in the post-period, the median DTN time was reduced significantly from 70 to 39 minutes, and a significantly higher number of patients were discharged home in the post-period (46.5% to 59.5%) (Kamal et al., 2020). Using data from 71,169 patients admitted to 1,030 GWTG-participating hospitals, the outcomes and process times of patients admitted before and after the initiation of a quality improvement initiative (Target:Stroke) were examined (Fonarow et al., 2014). During that time the median DTN were reduced significantly from pre- to post- intervention (77 vs. 67 minutes, p<0.001), the percentage of patients treated within 60 minutes of stroke onset increased significantly from 26.5% to 41.3%, and in-hospital mortality decreased significantly from 9.93% to 8.25%. The percentage of patients discharged home also increased significantly from 37.6% to 42.7%.

            The results from several studies indicate that tenecteplase, which has some pharmacokinetic advantages over alteplase, may be non-inferior to alteplase. Several clinical trials are ongoing and the results are not yet available. In these trials tenecteplase was compared with either alteplase (ATTEST2 NCT0281440) or placebo, or best medical management (TIMELESS NCT03785678, TWIST NCT03181360, and TEMPO-2 NCT02398656). Among completed trials comparing tenecteplase with alteplase, all were used as a potential bridging treatment prior to thrombectomy. The Alteplase Compared to Tenecteplase in Patients with Acute Ischemic Stroke (AcT) Trial (Mennon et al., 2022) was the first trial to report that tenecteplase is non-inferior to alteplase for 90-day functional outcomes. In this trial, 1,600 patients recruited from 22 centres who were eligible for treatment with alteplase (+/-thrombectomy) were randomized to receive intravenous tenecteplase (0.25mg/kg, maximum 25m) or 0.9 mg/kg alteplase. At a median of 97 days 36.9% of patients in the tenecteplase group achieved the primary outcome (mRS score of 0-1) vs.34.8% in the alteplase group (unadjusted difference=2.1%, 95% CI -2.6% to 6.9%; adjusted RR=1·1, 95% CI 1·0 to 1·2), meeting the non-inferiority threshold, (the lower bound 95% CI of which was set at >-5%). There was no significant difference between groups in mortality at 90 days (15.3% vs. 15.4%), or in the proportion with symptomatic ICH at 24 hours (3.4% vs. 3.2%). In contrast to these findings, the NOR-TEST 2 (Kvistad et al., 2022) was halted early due to safety concerns, which included an increased risk of intracranial hemorrhage and mortality; however, the dose in the tenecteplase group was higher (0.4 mg/kg) than is currently recommended (0.25 mg/kg). In the EXTEND-IA TNK (Campbell et al., 2018), which compared 0.25mg/kg tenecteplase vs. 0.9 mg/kg alteplase, at initial angiographic assessment, a significantly higher number of patients in the tenecteplase group achieved substantial reperfusion (22% vs. 10%, p=0.02 for superiority), although the percentage of patients who were functionally independent at 90 days or who had achieved an excellent outcome, did not differ between groups.

            Sex and Gender Considerations

            Possible interactions (treatment group x sex) were not analyzed in the initial reports of early trials of alteplase including NINDS (1995), ATLANTIS (1995), or ECASS (1995, 1998, 2008). The IST-III examined this relationship and reported there were no significant interactions based on sex, as did the authors of the ENCHANTED trial (2016) that examined low vs. standard dose alteplase. In more recent trials of late window treatment, including WAKE-UP (2018) and EXTEND (2019), the results of subgroup analyses based on sex were not conducted or reported. In the RCTs of tenecteplase including NOR-TEST 2 (2022), EXTEND-IA TNK (2018), and NOR-TEST (2017), the effect of treatment based on sex was not reported in subgroup analyses in the initial publications. Subgroup analysis for interactions based on sex for the AcT trial (2022) were conducted and no interactions were found.

            Endovascular Thrombectomy

            Re-vascularization can also be achieved through mechanical dislodgement with specialized devices (+/- intra-arterial and/or intravenous rt-PA). To date, over 10 major RCTs have been completed for which results have been published, in which endovascular therapies were compared with best medical management. Several trials are still ongoing or have yet to report their findings (e.g., TENSION NCT03094715; MR CLEAN-LATE ISRCTN 19922220). The recent results from most of these trials indicate that rapid endovascular therapy is a safe and more effective treatment than intravenous t-PA alone, for patients with anterior circulation ischemic strokes in selected regions, when performed within 6 to 24 hours of symptom onset.

            In the one of the earliest trials, MR CLEAN (Berkhemer et al., 2014), included 500 adult patients with a baseline NIHSS score of 2 or greater, and who were treatable within 6 hours of stroke onset. Patients were randomized to receive endovascular treatment with rt-PA or urokinase, and/or mechanical treatment with retrievable stents, which were used in 81.5% of patients, or other available devices, versus best medical management. The median time from stroke onset to groin puncture was 260 minutes. The majority of patients in both groups were treated with intravenous t-PA (87.1% intervention group, 90.6% control group). There was a significant shift in the distribution towards more favourable mRS scores among patients in the intervention group at 90 days (adj common OR=1.67, 95% CI 1.21-2.30). The odds of both a good (mRS 0-2) and excellent (mRS 0-1) recovery at day 90 were also significantly higher among patients in the intervention group (adj OR=2.07, 95% CI 1.07-4.02 and adj OR=2.16, 95% CIU 1.39-3.38, respectively). Patients in the intervention group were more likely to have a lower median final infarct volume (-19 mL, 95% CI 3-34, n=298). At two-year follow-up (van den Berg et al., 2017), the odds of an mRS score of 0-2 remained significantly higher in the intervention group (37.1% vs. 23.9%, adj OR= 2.21, 95% CI 1.30−3.73, p=0.003). The ESCAPE trial (Goyal et al., 2015) enrolled 316 patients ≥18 years, with stroke onset 5 and moderate-to-good collateral circulation. Patients were randomized to receive endovascular mechanical thrombectomy, using available devices or best medical management. The median time from stroke onset to first reperfusion was 241 minutes. 72.7% of patients in the intervention group and 78.7% of those in the control group received intravenous t-PA. The odds of improvement in mRS scores by 1 point at 90 days were significantly higher among patients in the intervention group (adj OR=3.2, 95% CI 2.0-4.7). The odds of good outcome (mRS score 0-2) at 90 days were also higher in the intervention group (adj OR=1.7, 95% CI 1.3-2.2), as were the odds of a NIHSS score of 0-2 and a Barthel Index score of 95-100 (adj OR=2.1, 95% CI 1.5-3.0 and 1.7, 95% CI 1.3-2.22, respectively). The risk of death was significantly lower in the intervention group (adj RR=0.5, 95% CI 0.-0.8). In neither MR CLEAN nor ESCAPE, was there an increased risk of symptomatic ICH associated with endovascular therapy. No interaction effects were found in subgroup analyses of age, stroke severity, time to randomization, or baseline ASPECTS in either of the trials.

            The results of the DAWN (Nogueira et al., 2018) and DEFUSE-3 (Albers et al., 2018) trials indicate that the treatment window for mechanical thrombectomy is wider than previously thought. The DAWN trial included 206 patients, last been known well 6 to 24 hours earlier, with no previous disability (mRS 0-1) and who met clinical mismatch criteria who had either failed intravenous t-PA therapy, or for whom its administration was contraindicated because of late presentation. Patients were randomized to treatment with thrombectomy with Trevo device plus medical management, or medical management alone. The trial was terminated early after interim analysis when efficacy of thrombectomy was established. The median intervals between the time that a patient was last known well and randomization was 12.2 hours in the thrombectomy group and 13.3 hours in the control group. The mean utility weighted mRS score was significantly higher in the thrombectomy group (5.5 vs. 3.4, adj difference =2.0, 95% Cr I 1.1-3.0, prob of superiority >0.999). There were no interactions in subgroup analysis (mismatch criteria, sex, age, baseline NIHSS score, occlusion site, interval between time that patient was last known well and randomization and type of stroke onset). A significantly higher proportion of patients in the thrombectomy group experienced an early response to treatment, had achieved recanalization at 24 hours, and were independent (mRS 0-2) at 90 days (49% vs. 13%, NNT=3). The admission criteria for the DEFUSE-3 trial were broader and included those who had remaining ischemic brain tissue that was not yet infarcted. The median time from stroke onset to randomization was just under 11 hours for patients in the endovascular group. A significantly higher proportion of patients in the endovascular group were independent (mRS 0-2) at 90 days (45% vs. 17%, OR=2.67, 95% CI 1.60–4.48, p

            The results from several meta-analyses that pooled the results from these major trials indicate better outcomes were associated with mechanical thrombectomy. A recent Cochrane review (Roaldsen et al., 2021) included the results from 19 RCTs (3,793 participants). Treatment with EVT was associated with a significantly higher likelihood of favourable outcome (RR=1.61, 95% CI 1.42 to 1.82) with a high certainty of evidence, without a significantly increased risk of symptomatic ICH (RR=1.46, 95% CI 0.91 to 2.36). A pooled analysis of 6 RCTs including DAWN, DEFUSE 3, ESCAPE, RESILIENT, POSITIVE and REVASCAT also found significantly better outcomes in patients in the EVT group (Jovin et al., 2021). The was a significant shift in the ordinal analysis of mRS scores favouring less disability in the thrombectomy group (adj OR=2·54, 95% CI 1·83–3·54) with stronger effects noted with earlier treatment.

            For large artery occlusions in the posterior circulation, the data are limited with the results from two RCTs in which EVT was compared with best medical management. Liu et al. (2020) included 131 adult patients presenting within 8 hours of vertebrobasilar occlusion to 28 centres in China in the Basilar Artery Occlusion Endovascular Intervention versus Standard Medical Treatment (BEST) trial. The trial was terminated early due to excessive crossovers and low enrollment. In the intention-to-treat analysis, the percentage of patients with a favourable outcome (mRS 0-3) at 90 days was not significantly higher in the intervention group (42% vs.32%; adjusted [age and baseline NIHSS] OR=1·74, 95% CI 0·81–3·74, p=0.23), nor was the percentage of patients who were functionally independent; however, in both the per protocol and as treated analyses, the percentage of patients with a favourable outcome was significantly higher in the EVT group. There was no significant difference between groups in 90-day mortality or in symptomatic intracerebral hemorrhage (sICH). The Basilar Artery International Cooperation Study (BASICS) trial recruited 300 patients with basilar artery occlusion (Langezaal et al., 2021). Intravenous alteplase was used in close to 80% of patients in the EVT and control groups. The percentage of patients in the EVT group who experienced a favourable (mRS 0-3) or excellent (mRS 0-2) outcome at 90 days was not significantly higher in the EVT group. The results of these two RCTs and three observational studies were pooled in a systematic review by Katsanos et al. (2021). With low certainty of evidence, there was no significant difference found between the groups for the primary outcome of mRS score 0-3 at 90 days (RR= 0.97, 95% CI: 0.64-1.47). There were no significant differences between groups for the proportion of patients with mRS scores of 0-2 at 3 months, all-cause mortality or functional outcome (shift analysis), with significant heterogeneity. The risk of sICH was significantly higher in the EVT group (RR=5.42, 95%CI: 2.74-10.71).

            To date, 6 RCTs have been published comparing direct EVT with intravenous alteplase prior to EVT (i.e., bridging). The most recent trials, SWIFT DIRECT (Mitchell et al., 2022) and SWIFT DIRECT (Fischer et al., 2022) both reported that EVT alone was not shown to be non-inferior to EVT plus thrombolysis. For the primary outcome of mRS score of 0-2 at 90 days, the adjusted differences in proportions between groups were −7·3% (95% CI −16·6 to 2·1, p=0·12) in the SWIFT-DIRECT trial and −5·1% (95% CI −16 to 5·9, p=0·19) in the DIRECT SAFE trial, which crossed the lower boundaries of the two-sided 95% confidence interval set for non-inferiority at 12% and 10%, respectively. Two previously published trials, the SKIP trial (Suzuki et al., 2021) and MR CLEAN–NO IV trial (LeCouffe et al., 2021), also did not demonstrate the non-inferiority of EVT alone. In the MR CLEAN–NO IV trial, the adjusted common odds ratio (OR) for shift in mRS score at 90 days was 0.84 (95% CI 0.62 to 1.15), which showed neither superiority nor non-inferiority of EVT alone. In the SKIP trial, mechanical thrombectomy alone was not associated with a favourable shift in the distribution of the mRS score at 90 days (OR=0.97, 1-sided 97.5% CI, 0.60 to ∞; non-inferiority p = .27, which crossed the 0.74 threshold). In contrast, DIRECT-MT (Yang et al., 2021) and DEVT (Zi et al., 2021) reported that EVT alone was non-inferior to alteplase followed by EVT. In the DEVT trial, 54.3% of patients in the EVT group achieved functional independence vs. 46.6% in the bridging group (difference= 7.7%; 1-sided 97.5% CI, −5.1% to ∞; p = .003 for non-inferiority, threshold for non-inferiority was -10%). Finally, EVT alone was non-inferior to bridging in an ordinal shift analysis of mRS scores at 90 days (adjusted common OR=1.07; 95% CI 0.81 to 1.40; p=0.04 for non-inferiority) in the DIRECT-MT trial.

            There have been a limited number of RCTs specifically comparing the use of general anesthesia versus conscious sedation for EVT procedure. Preliminary results of one of the most recent trials the Anesthesia Management in Endovascular Therapy for Ischemic Stroke (AMETIS) trial (NCT03229148), indicate that conscious sedation increased the probability of a good outcome (mRS 0-2) at 90 days by 29%. Previous single-centre trials including the General or Local Anesthesia in Intra Arterial Therapy (GOLIATH, Simonsen et al., 2018), Anesthesia During Stroke (AnStroke Trial, Löwhagen Hendén et al., 2017) and Sedation vs. Intubation for Endovascular Stroke Treatment (SIESTA, Schönenberger et al., 2016), reported that general anesthesia was associated with better outcome (mRS 0-2) at 90 days. The conversion from conscious sedation to general anesthesia in these trials occurred in 6.3%, 14.3%, and 15.5%. The results of the SEdation Versus General Anesthesia for Endovascular Therapy in Acute Ischemic Stroke (SEGA, NCT03263117) are expected in early 2023. A systematic review including the results of all three aforementioned trials plus data from a pilot study of 40 patients (Campbell et al., 2021) found the odds of successful recanalization and good functional outcome were significantly higher in the general anesthesia group (OR=2.14, 95% CI 1.26-3.62, p=0.005 and OR=1.71, 95% CI: 1.13-2.59; P=0.01, respectively), with no significant differences between groups in the risk of mortality or intracerebral hemorrhage. A Cochrane review (Tosello et al. 2022) included the results from 7 RCTs and reported on both short and long-term outcomes. In the short-term, general anesthesia was not associated with better early neurological recovery or stroke related mortality, but was associated with a decreased risk of adverse events and greater likelihood of artery revascularisation. The likelihood of having a good functional outcome (mRS ≤2) at 90 days was not significantly greater in the general anesthesia group. The outcomes of patients who received general anesthesia or conscious sedation has also been examined in the context of EVT RCTs. Using the results from 7 RCTs including MR CLEAN, ESCAPE, EXTEND-IA, SWIFT PRIME, REVASCAT, PISTE and THRACE, Campbell et al. (2018) performed a patient-level meta-analysis comparing the outcomes of patients randomized to the mechanical thrombectomy groups who had received general anesthesia or non-general anesthesia. The odds of improved outcome using non-general anesthesia were significantly higher in ordinal analysis of mRS scores. The authors estimated for every 100 patients treated under general anesthesia (compared with non-general anesthesia), 18 patients would have worse functional outcome, including 10 who would not achieve functional independence. There was no increased risk of 90-day mortality associated with general anesthesia.

            Sex and Gender Considerations

            In a patient-level meta-analysis using data from 5 RCTs, conducted by the HERMES Collaborators (2016), there were no significant treatment effects of EVT based on pre-specified subgroups including age, sex, NIHSS, site of intracranial occlusion, intravenous alteplase received or ineligible, ASPECTS, time from onset to randomisation, or the presence of tandem cervical carotid occlusion. The same finding was reported in another Hermes Collaboration, using data from 7 RCTs (Chalos et al. 2019), which was confined to an examination of sex differences. Subgroup analyses based on sex were not conducted for the primary outcome in the RESCUE-Japan LIMIT (2022) or REVESCAT (Jovin et al. 2015) trials, although other potential effect modifiers (e.g. age, baseline NIHSS score) were examined. No evidence of heterogeneity of treatment effect based on sex was detected in prespecified subgroups in the DEFUSE 3 trial (Albers et al. 2018), DAWN trial (Nogueira et al. 2018), ESCAPE (Goyal et al. 2015), or THRACE (2016) trial, where subgroup analysis was performed. In the two, most recent trials examining direct EVT with bridging therapy, (DIECT SAFE, Mitchell et al. 2022 and SWIFT DIRECT, Fischer et al. 2022), differences in sexes between treatment groups were examined in prespecified subgroup analyses; none were found. Sex differences were examined specifically in 3,422 patients included in the IRETAs database who had undergone EVT treatment since 2011 (Casetta et al. 2022). The outcomes of women vs. men were compared in the original cohort (1,621 men and 1,801 women) and in a propensity-matched cohort of 1,150 men and women. In both the whole cohort and matched-pair cohort, the odds of functional independence at 90 days given EVT treatment were significantly higher in women (OR= 1.19, 95% CI 1.02–1.38 and OR=1.25, 95% CI 1.04-1.51, respectively). Time metrics (e.g., onset to groin puncture) were similar for men and women.

            Note: The CSBPR Acute Stroke Management writing group and the National Advisory Committee strongly endorse all of the recommendations in Section 5, based on available research evidence, clinical expertise, and international consensus. A recent technology assessment report provided a focused assessment of some of these data and suggested there is “substantial uncertainty” regarding the effectiveness of alteplase; however, this technology report did not synthesize all the available evidence and their conclusions differ from most other international guideline organizations as well as the CSBPR writing group. Refer to evidence table (CADTH, 2022).

            Stroke Resources

            Previous Section 4. Emergency Department Evaluation and Management of Patients with Acute Stroke and TIA Next Section 6. Acute Antithrombotic Therapy