Health Technology Assessment
Mechanical thrombectomy for acute ischemic stroke
Health technology assessment|
The Commissioner Forum for the Regional Health Authorities («Bestillerforum RHF») requested a HTA on the use of mechanical thrombectomy compared to standard treatment for the treatment of acute ischemic stroke.
Each year around 10,000 patients are treated for severe acute stroke at Norwegian hospitals. Although many patients die, survivors often suffer from considerable loss of function. Mechanical thrombectomy is a treatment option for large vessel ischemic stroke. Thrombectomy consists in physically removing the obstructing clot (thrombus) using specific for-purpose designed devices. The method is used when standard treatment (thrombolysis) has failed to dissolve the thrombus, or is contra-indicated. In Norway, thrombectomy is performed at specific intervention centres. However, time is a crucial factor, as the intervention is recommended within maximum 6 hours following the stroke. Thus implementation of this method at the national level, i.e. being available to all citizens, will meet important challenges due to the particular population distribution, topography and weather conditions in large parts of our country.
Clinical effect and safety:
Available research documentation on the use of mechanical thrombectomy as additional treatment to standard therapy (thrombolysis) compared to standard therapy alone shows that:
- No certain difference between the two therapeutic options in terms of mortality assessed after 90 days
- The risk of functional loss is decreased with thrombectomy
- Health-related quality of life assessed after 65-90 days is probably better following thrombectomy, but the quality of evidence is low
- Assessed 24-30 hours after treatment, there might be increased risks of haemorrhages (both symptomatic and non-symptomatic ones) following thrombectomy, however there is no significant difference when symptomatic intracranial haemorrhages are analysed separately,
- Risks for recurrent strokes at 90 days are possibly higher after thrombectomy, but this is uncertain due to low quality of the evidence,
- No clear cut conclusions can be drawn in terms of adverse effects related to thrombectomy.
- There are arguments for both a decentralised and a centralised health care organisational model for acute stroke care depending on whether focus is on thrombolysis or thrombectomy, and local conditions.
- If the intention is to provide equal stoke care, establishing more intervention centres, developing pre-hospital health services and collaboration between regions, as well as neighboring countries should be considered.
- Costs associated with thrombectomy as additional treatment to thrombolysis covering procedure, diagnostical imaging and transport are estimated to 84,331 NOK (range: 51,631 – 192,631 NOK).
Important ethical issues are related to:
- The principle of access to equal health care services
- Patients’ autonomy, particularly when the stroke has led to decreased decision-making capacity
According to records from the National Heart and Cardiovascular Registry for 2014, around 11,000 patients with acute severe stroke were treated at Norwegian hospitals, and 85% were ischemic caused by occlusion of arteries supplying the brain. Endovascular therapy using mechanical thrombectomy for acute ischemic stroke physically removes the clot (thrombus) occluding the artery. This method allows larger thrombi not dissolved by thrombolysis to be removed from proximal anterior artery branches in the brain, and provides a therapeutic alternative to patients contra-indicated for thrombolysis.
The national guidelines for stroke care are under revision, and whether mechanical thrombectomy on selected patients with occlusions in large brain arteries should be offered routinely, is currently being considered.
The Commissioner Forum for the Regional Health Authorities («Bestillerforum RHF») has asked the Knowledge Centre for the Health Services in the Norwegian Institute of Public Health to carry out a HTA on the use of mechanical thrombectomy compared to standard treatment for the treatment of acute ischemic stroke.
In collaboration with a group of experts within the field, we have performed an HTA that includes the following:
- A systematic review of clinical effect and safety on the use of mechanical thrombectomy in addition to standard treatment for acute ischemic stroke compared to standard treatment alone
- An assessment of organisational consequences of introducing thrombectomy as treatment option at the national level
- An economic evaluation including a overview of relevant economic assessments on thrombectomy for the treatment of acute ischemic stroke, along with an estimate of costs related to the treatment in Norway
- Ethical issues related to thrombectomy as routine treatment option
In addition to costs associated the different treatments options for acute ischemic stroke, we expect costs to vary according to the different organisational models for mechanical thrombectomy and the pre-hospital health care. As the European HTA we are using in our HTA identifies a clinical effect of thrombectomy, we will perform a cost utility analysis (CUA). This will be published as part two of this HTA within summer 2016. It will be based on the information gathered in the «Organisation» chapter included in this first part of our report.
The aim of this HTA is to answer following research questions:
1) What is the effect of mechanical thrombectomy for the treatment of patients with acute ischemic stroke with regard to:
- Mortality at 90 days
- Function assessed with the modified Rankin scale (mRS) at 90 days
- Safety including side-effects and adverse events
- Health-related quality of life (HRQoL)
2) What are the organisational preconditions and further consequences of introducing thrombectomy as a treatment option at the national level in Norway?
3) What are the costs associated with thrombectomy?
4) What are the ethical implications related to thrombectomy introduced as a part of routine treatment in Norway?
Our assessments are based on the use of other systematic reviews when possible. In case no previously performed systematic review related to our research question has been done or does not meet our criteria for methodological quality, we perform our own search for primary literature.
At the very beginning of this project, in May 2015, we identified a newly initiated HTA by the EUnetHTA Collaboration, which was relevant to our research question.
When this European HTA was published December 2015, we considered the quality to be high, according to our criteria for systematic reviews. Thus, we chose to use this HTA for the assessment of clinical effectiveness and safety.
For organisational issues, we have together with the expert group and other stakeholders mapped the treatment chain in Norway for acute ischemic stroke with focus on patients who are prone to benefit from thrombectomy. We have looked into the current organisation, and have pointed to some of the main organisational challenges related to the introduction of the method in the clinical routine of stroke health care. We have also examined other HTAs that have dealt with organisational issues, official investigations and reports, scientific articles specifically dealing with the organisational challenges, and data on demographic, topographic and geographic settings. Information from this organisation chapter will constitute a major basis for the further economic analyses, and provide important background information on the organisation of this treatment alternative in Norway.
For the economic evaluation, we have first searched for published economic evaluations and communicated these. Further, we have performed an evaluation of costs based on information obtained from the Norwegian Health Directorate and hospital trusts.
The aim of including ethics in an HTA is to raise awareness, inform and discuss ethical aspects and challenges related to the introduction of the method into the Norwegian health services. For assessing of ethical issues we have used our check-list elaborated for the purpose in addition to consulting other HTAs where ethical aspects have been discussed.
Clinical effectiveness and safety
The report from the EUnetHTA Collaboration published December 2015 included eight RCTs (three from 2013 and five from 2015) with totally 2,423 patients. The studies compared standard treatment, i.e. intravenous thrombolysis (when not contra-indicated) with standard treatment combined with mechanical thrombectomy. The five RCTs from 2015 including 1,287 patients were the ones relevant for the current Norwegian stroke care setting, as these have used so-called stent-retrievers (last generation), which is the equipment used in Norway today. These studies were included in sub-group meta analyses when possible, and our conclusion were basically based on these.
Mortality was reported at 90 days in all the included studies. Meta analysis of the five RCTs from 2015 favoured thrombectomy, but the effect was not statistically significant, and estimates pointed in both directions. RR was 0.82 with 95% CI 0.60 to 1.11 (GRADE-level low).
According to the meta analysis of the five more recent studies, likelihood of obtaining independent function (set at mRS ≤2 at 90 days) was significantly higher with thrombectomy compared to standard treatment. RR was 1.72 with 95% CI 1.03 to 1.80 (GRADE-level moderate).
Level of function using the Barthel ADL Index showed significantly better results with thrombectomy according the meta analysis of three of the RCTs from 2015 (n=938). RR was 1.70 with 95% CI 1.45 to 2.01 (GRADE-level moderate).
Seven of the included studies had assessed outcomes related to reperfusion. Although thrombectomy showed better results than standard treatment alone, the evidential basis was uncertain, because measurements had been done at different time points, using different methods, or done in different ways (e.g. by dividing into groups of patients according to results or by doing various forms of dichotomisation). Studies were therefore not included in any meta analyses, and no definitive conclusions could be drawn. The same applied for the degree of reperfusion using mTICI score assessed at final angiography.
Three of the studies from 2015 (n=1,021) assessed health-related quality of life (HRQoL) using the EQ-5D questionnaire. Results from all three studies were in favour of thrombectomy, however different versions of the questionnaire and different scales had been used in the various studies.
The same eight RCTs were basis for assessing safety issues. Additional six studies were included (two RCTs and four observational studies published 2012-2013 with totally 641 patients), as these had reported device-related adverse effects. Same limitations did not allow meta analyses for some of the clinical effectiveness outcomes applied for several safety outcomes.
According to meta analysis of the five RCTs from 2015, there were significant more cerebral haemorrhages (both symptomatic and asymptomatic ones) assessed after 24-30 hours with thrombectomy compared to standard treatment. RR was 1.46 with 95% CI 1.07 to 1,99 (GRADE-level low), but the implication of this result is difficult to interpret, as all cases of haemorrhages were reported regardless of clinical significance. Meta analysis for symptomatic intracranial haemorrhages (SICH) did however not show any difference between the treatments, and pointed in both directions. RR was 1.08 with 95% CI 0.64 to 1.83 (GRADE-level low).
Recurrent ischaemic strokes within 90 days were reported in three of the RCTs from 2015 (n=1,021). The total effect estimate from the meta analysis favoured standard treatment, but was not statistically significant. RR was 3.09 with 95% CI 0.86 to 11.11 (GRADE-level low).
It was difficult to get any clear picture from the seven RCTs reporting on adverse effects. Five of the studies had gathered all types of adverse effects, which occurred in 10.9 to 29.1% of the patients treated with thrombectomy. Only one of the RCTs specifically reported that 5.1% of the patients had encountered device-related adverse events. Six other studies (those included additionally for the assessment of safety) reported 2.8 to 13.5% of patients with device-related adverse events.
None of the included studies had recorded harms due to the procedure. Intra- and inter-observational variations in terms of interpreting findings or user-related errors related to the procedure were not assessed either.
Establishing thrombectomy as a treatment option at national level will lead to changes in current clinical routines and require investments in new facilities. The method implies development of competence both in pre-hospital health service and at hospitals, and the implementation will affect distribution of functions and coordination between different levels of health services.
The time window between the appearance of first stroke symptoms and completed thrombectomy can be no more than 6 hours. Time is therefore of major importance for the result of the treatment. Due to its demography, topography and weather conditions, Norway will meet important organisational challenges if the treatment option is to be offered to the whole population. All treatments for stroke aimed at re-establishing blood circulation are urgent. Thus conflicting considerations and interests may arise between the need to organise the thrombolytic treatment and general stroke care as near as possible to where the patient lives in a decentralised model, and the need for centralising high-specialised care such as endovascular treatment. Taking into account both of these concerns in a way that both provides equal access to efficient treatment independently of where patients are living, and at the same time ensures high quality treatment, is challenging.
The need for developing the pre-hospital health services, and in particular, ambulance transport, and the need of establishing more intervention centres must be considered. Solutions to ensure competence and experience of those performing the treatment must be generated. Various forms of collaboration between health regions, but also with neighbouring countries should be investigated. In addition, the organisation of this treatment option should be considered in relation to the treatment of comparable health conditions.
How the above mentioned challenges should be met, must be further resolved at regional and national levels, and adapted to local settings.
Costs associated with thrombectomy as additional treatment to thrombolysis for an average patient in acute phase, covering procedure, diagnostic imaging and transport are estimated to 84,331 NOK (range: 51,631 – 192,631 NOK).
New methods may challenge moral norms and values, and introducing these into health care may imply reallocation of resources. Ethical issues related to the introduction of thrombectomy routinely into clinical practice in Norway, are mainly associated to the aim of providing equal access to the treatment independent of where people lives. The question regarding which patients should be offered the treatment when considering age, comorbidity, etc. is also important. Further, ethical dilemmas are related to patient’s autonomy, especially when the stroke has led to decreased decision-making capacity. Ethical concerns related to introducing thrombectomy routinely will have consequences for how resources will be prioritised between new treatment options and current practice. This last issue will be discussed more in depth in the second part of this HTA, which comprises the cost-utility analysis.
Results on clinical effect and safety should be considered in light of differences among the included studies. Different variants of devices were used to remove the thrombus. Moreover, various approaches were used to select patients for thrombectomy, both regarding diagnostic imaging and clinical assessment. Proportion of patients treated with thrombolysis before thrombectomy varied among the included studies. However a study from February 2016 by Goyal and collaborators have shown that the effect size on the mRS score was equally distributed between subgroups of patients, as for instance between patients who had received thrombolysis before thrombectomy and those who had not, between different age groups and between different types of occlusions.
The main elements of the cost estimate we have provided comes from the Clinic for Radiology and Nuclear Medicine (KRN) at the Oslo University Hospital (OUS). Costs may therefore differ to some extent from other hospital trusts. We believe however that numbers received from KRN present a balanced picture of price levels for the various procedures, and that they can be used across hospital trusts, and thus work well as estimates. The price of the procedure package is based on the current capacity at OUS, but costs related to establishment of a new centre is not part of our estimate.
Available documentation indicates that mechanical thrombectomy for treating selected patients suffering from acute large-vessel ischemic stroke has a positive effect on function, morbidity and health-related quality of life, but there is no significant difference between thrombectomy and standard treatment with regard to mortality. There is possibly an increase of risk for brain haemorrhages after thrombectomy when they look at both symptomatic and asymptomatic haemorrhages, but there is no certain difference when it comes to serious symptomatic haemorrhages. The risk of encountering a new stroke within 90 days is possibly higher with thrombectomy, but there is not enough evidence to definitively confirm this. There is also insufficient evidence to draw any clear conclusions on adverse events.
Assessing the organisational alternatives for the possible introduction of thrombectomy as a part of routine treatment options for acute stroke, will require thorough examination of local settings and mapping of available and needed resources, in addition to loco-regional demographic and topographic situations, hospital structures, and possibilities of collaboration between health regions and neighbouring countries. Since the time factor is crucial, pre-hospital services are central. In addition, trade-offs between the need for a decentralised organisational model when it comes to the overall acute stoke care including thrombolysis, and the need for centralised high-specialised health care services, such as endovascular stroke therapy, have to be made.
Additional costs associated with thrombectomy as supplementary treatment to thrombolysis for an average patient in acute phase covering expenses for the procedure, diagnostic imaging and transport is estimated to around 81 500 NOK (range: 48 756 – 189 756 NOK).
We would like to thank the authors of the rapid assessment «Endovascular therapy using mechanical thrombectomy devices for acute ischaemic stroke» produced by the European network for HTA (EUnetHTA) which our assessment has been based on for assessing clinical effectiveness and safety.