Loading prototype

Health Technology Assessment

Health Technology Assessment

Patent foramen ovale closure, antiplatelet therapy or anticoagulation therapy alone for management of cryptogenic stroke

YEAR
2019

AUTHORS
Hagen G, Huitfeldt A, Vandvik PO

ISBN (DIGITAL)
978-82-8082-993-1

Executive summary and conclusions, in light of key prioritization criteria

Patients with cryptogenic stroke due to patent foramen ovale (PFO) have high risk of recurrent strokes with currently recommended treatment (antiplatelet therapy) (moderate severity condition). In patients below 60 years PFO closure probably confers an important reduction in ischemic stroke recurrence compared with antiplatelet therapy alone (8%) but may make no difference compared with anticoagulation. PFO closure incurs a risk of persistent atrial fibrillation (2%) and device-related adverse events (3.6%). Compared with alternatives, anticoagulation probably increases major bleeding. Cost-effectiveness was demonstrated for PFO closure compared to antiplatelet therapy and anticoagulation. Organisational aspects include capacity of PFO closure at regional hospitals and budget implications.

Question 1: PFO closure or antiplatelet therapy?

Conclusion: Moderate to high certainty evidence demonstrates that PFO closure confers an 8% absolute reduction in stroke, no important differences in death or major bleeds, a 2% increase in atrial fibrillation and 3.6% increase in device related complications. PFO closure was cost-effective (NOK -505 330 per QALY).
Surgical closure of PFO will compared to medical management with antiplatelet therapy, result in health gains of 1.49 Quality-Adjusted Life-Years (QALY).

Comparison of benefits and harms

Incremental cost: Over a lifetime perspective, PFO closure will result in a cost saving of NOK 755 339 per patient as compared to treatment with antiplatelet therapy.

Budget impact: A national introduction of PFO closure will, if all new cases are treated, have an estimated budget impact of 34 mill NOK per year.

Patent Foramen Ovale will likely lead to an absolute shortfall of 14.8 life years in good health (QALYs) over a lifetime perspective.
Health gains Incremental costs Main ICER
1.49 QALYs
Will be gained with PFO closure resulting in a substantial increase in health.
NOK -755 339
Based on the assumption of extended indication for PFO closure.
NOK -505 330
Average incremental cost-effectiveness ratio per patient for PFO closure.
Intervention Costs (NOK) Incremental costs (NOK) Effects (QALYs) Incremental effect (QALYs) ICER (NOK/QALY)
Antiplatelet treatment 2 925 387 9.4753
PFO closure 2 170 048 -755 339 10.9701 1.4947 -505 330

The results show that the total expected average costs per patient in a lifetime perspective were 2 925 387 NOK for patients who get antiplatelets and 2 170 048 NOK for patients who undergo PFO closure. The incremental cost for PFO closure patients is thus about -755 339 Norwegian kroner. At the same time, the PFO closure patients are in better health, with an average difference of 1.4947 QALYs. The incremental cost per additional QALY (ICER) () is -505 330 Norwegian kroner per QALY gained.

Organisational aspects

Current situation: PFO closure is already offered to some patients with cryptogenic stroke and PFO in two university hospitals in Norway (Haukeland and Rikshospitalet), with approximately 150 procedures per year. There are no national guidelines for this procedure and no systematic processes for referral from hospitals treating patients for stroke.
Future situation: If PFO closure will be approved for use in Norway there will be a need to organise health services at the central and local level, according to updated national stroke guidelines planned by the Norwegian Directorate of Health spring 2019. If the national guidelines end up with a strong recommendation to offer PFO closure (as did the BMJ Rapid Recommendations panel) it implies that the large majority of patients should undergo PFO closure. The organisational consequences would then be as follows:

University hospitals need to have sufficient capacity and expertise to perform the procedures. The anticipated need is 400 procedures per year, for newly diagnosed patients. If PFO closure is going to be systematically offered to patients previously diagnosed with cryptogenic stroke and PFO closure the number of procedures will be substantially higher (e.g. 1000 procedures per year). The experts who contributed in the health technology assessment recommends that St.Olavs Hospital and UNN also should offer PFO closure to ensure sufficient capacity and equal access to this procedure.

All hospitals who manage patients with stroke need to ensure adequate diagnosis (to identify patients with cryptogenic stroke and PFO) and referral for PFO closure. This requires that clinicians refer patients with cryptogenic stroke below 60 years of age to cardiological assessment with transesophageal eccocardiography (TEE) to determine if PFO is present or not. Cardiologists then need expertise and capacity to perform TEE with satisfactory quality. Whereas we believe all hospitals have cardiologists and equipment for TEE it remains uncertain to what extent this service could be offered with satisfactory quality at local hospitals. An alternative would be to refer patients to the university hospitals for diagnostic assessment of PFO.

Question 2: PFO closure or anticoagulation therapy?

Conclusion: Low certainty evidence for no important difference in recurrent ischemic stroke, death, moderate certainty evidence for increased major bleeds with anticoagulation (2%), with PFO closure resulting in a 2% increase in atrial fibrillation and 3.6% increase in device related complications. PFO closure was potentially cost-effective (NOK -637 195 per QALY).
Surgical closure of PFO will compared to medical management with anticoagulation, result in health gains of 0.58 Quality-Adjusted Life-Years (QALY).

Comparison of benefits and harms

Incremental cost: Over a lifetime perspective, PFO closure will result in a cost saving of NOK 755 339 per patient as compared to treatment with antiplatelet therapy.

Budget impact: A national introduction of PFO closure will, if all new cases are treated, have an estimated budget impact of 34 mill NOK per year.

PFO will likely lead to an absolute shortfall of 14.8 life years in good health (QALYs) over a lifetime perspective.
Health gains Incremental costs Main ICER
0.59 QALYs
Will be gained with PFO closure.
NOK -379 663
Based on the assumption of extended indication for PFO closure.
NOK -637 195
Average incremental cost-effectiveness ratio per patient for PFO closure.
Intervention Total costs (NOK) Incremental costs (NOK) Effects (QALYs) Incremental effect (QALYs) ICER (NOK/QALY)
Anticoagulants 3 035 433 9.4862
PFO closure 2 655 770 -379 663 10.0820 0.5958 -637 195

The results show that the total expected average costs per patient in a lifetime perspective were 3 035 433 NOK for patients who get anticoagulants and 2 655 770 NOK for patients who undergo PFO closure. The incremental cost for PFO closure patients is thus about -379 663 Norwegian kroner. At the same time, the PFO closure patients are in better health, with an average difference of 1.4947 QALYs. The incremental cost per additional QALY (ICER) is -637 195 Norwegian kroner per QALY gained.

Organisational aspects

Current situation: PFO closure is already offered to some patients with cryptogenic stroke and PFO in two university hospitals in Norway (Haukeland and Rikshospitalet), with approximately 150 procedures per year. There are no national guidelines for this procedure and no systematic processes for referral from hospitals treating patients for stroke.
Future situation: If PFO closure will be approved for use in Norway there will be a need to organise health services at the central and local level, according to updated national stroke guidelines planned by the Norwegian Directorate of Health spring 2019. If the national guidelines end up with a strong recommendation to offer PFO closure (as did the BMJ Rapid Recommendations panel) it implies that the large majority of patients should undergo PFO closure. The organisational consequences would then be as follows:

University hospitals need to have sufficient capacity and expertise to perform the procedures. The anticipated need is 400 procedures per year, for newly diagnosed patients. If PFO closure is going to be systematically offered to patients previously diagnosed with cryptogenic stroke and PFO closure the number of procedures will be substantially higher (e.g. 1000 procedures per year). The experts who contributed in the health technology assessment recommends that St.Olavs Hospital and UNN also should offer PFO closure to ensure sufficient capacity and equal access to this procedure.

All hospitals who manage patients with stroke need to ensure adequate diagnosis (to identify patients with cryptogenic stroke and PFO) and referral for PFO closure. This requires that clinicians refer patients with cryptogenic stroke below 60 years of age to cardiological assessment with transesophageal eccocardiography (TEE) to determine if PFO is present or not. Cardiologists then need expertise and capacity to perform TEE with satisfactory quality. Whereas we believe all hospitals have cardiologists and equipment for TEE it remains uncertain to what extent this service could be offered with satisfactory quality at local hospitals. An alternative would be to refer patients to the university hospitals for diagnostic assessment of PFO.

Introduction

In this report, we evaluate the clinical effectiveness, safety, cost-effectiveness and organisational consequences of PFO closure in patients with cryptogenic stroke and PFO. This intervention aims to reduce the risk of recurrent stroke.

12 000 Norwegians have a stroke annually and represents an important cause of loss of life expectancy and quality of life (2). Stroke occurs most commonly in the older population, but approximately 20% patients are under the age of 60. Patients with a previous stroke are at increased risk of secondary strokes (3-5). Recurrent strokes are associated with a higher risk of cognitive worsening, drop-out of working life, problems with child care, loss of independence and death (6;7), compared to primary strokes.

Prevention of recurrent stroke aims to target the causes of the primary stroke. The recommended diagnostics for identifying possible aetiologically relevant factors includes blood-samples, ultrasound of the carotid and vertebral arteries, transcranial ultrasound, transthoracic and transoesophageal echocardiography and long-time heart rhythm monitoring. When these tests do not find any clear etiology, the stroke is classified as «cryptogenic». Approximately a third of all ischemic strokes are cryptogenic among young ischemic stroke patients. It is believed that some of the most common causes of cryptogenic strokes are paradoxical embolism (=embolic strokes) due to a right-to left shunt, most often communication between the right and left side of the heart (such as patent foramen ovale), less often located in the lungs or in other place.

Foramen ovale is an opening between the right and the left atrium of the heart, which has an embryological function in allowing the circulation to bypass the immature lungs by direct shunting of oxygen-loaded blood through the foramen ovale to the left heart, whitch then pump this blood to the whole body prior to birth. Normally, this opening closes at birth, when oxygen exchange via the lungs becomes possible. In approximately 25% of the population, however, the foramen ovale is not completely closed («patent foramen ovale=PFO»). This can allow blood clots that form in veins to bypass the lungs and travel into the systemic circulation, where they can cause a cerebral infarction or damage to other organs.

It is known that the prevalence of PFO is higher in patients who have had a cryptogenic stroke, than in people without stroke at the same age. While this suggests that the condition may play an etiological role in at least some patients, one must also consider the fact that 25% of the population have PFO, and that in many patients, the PFO is an incidental finding that must be interpreted in the context of other possible causes of stroke. Factors that suggest PFO as a cause of the stroke include a large PFO, with a big right-left shunt and the combination of PFO and an atrial septal aneurysm. Factors that suggest other causes than PFO, include any other reason for increased risk of stroke, such as atrial fibrilation. Patient with cryptogenic stroke and PFO are currently mostly treated with antiplatelet drugs (e.g. aspirin, dipyridamole or clopidogrel), according to national clinical practice guidelines (8). Some patients are also likely being treated with anticoagulation therapy (warfarin or direct oral anticoagulation treatment; DOAC).

Closure of PFO through implantation of a closure device via a catheter, is a mode of treatment which has become increasingly available. Such treatment is usually provided by interventional cardiologists. The known risks of PFO closure include procedure related complications as bleeding complications, pericardial effusion, perforation, embolic events, device embolization and atrial fibrillation and in the long run atrial fibrillation, endocarditis and erosion (9).

In August 2017, 3 high quality randomized trials were published in New England Journal of Medicine, comparing PFO closure with medical management with antiplatelet therapy or anticoagulation therapy. These trials, together with a more recent trial published March 2018, hold the potential to change clinical practice. This was reflected in strong recommendations for PFO closure published August 2018 (1), which triggered the request for the health technology assessment reported here. Similar recommendations have been issued in Denmark and Sweden during 2018.

Methods

Clinical-effectiveness

As proposed in the request for this health technology assessment by the Norwegian Stroke Foundation, we aimed to use a recently published systematic review of high quality, rather than duplicating evidence synthesis to assess clinical effectiveness of PFO closure. We were notified that surch a recent systematic review was published (10), linked to the BMJ Rapid Recommendation on PFO closure, August 2018 (1). In order to ensure identification of the most relevant and high-quality systematic review we performed a systematic literature search and selection process, based on the clinical question we formulated, inclusion outlined below.

Study type:
Systematic reviews of randomized trials
Population:
Patients with cryptogenic stroke (cerebral infarction) and patent foramen ovale
Intervention:
Closure of patent foramen ovale with any closure device (with or without antiplatelet therapy)
Comparator:
Antiplatelet therapy (Acetylsalicylic acid, Clopidogrel, etc)
Anticoagluation therapy (Warfarin, DOAC)
Outcome:
Recurrent stroke
Death
Transient ischemic attack
Major bleeding
Transient atrial fibrillation
Persistent atrial fibrillation
Pulmonary embolism
Systemic embolism
Device or procedure related complications
Languages:
Any languages

Literature search

Our research librarian (IH) planned and executed all systematic searches in collaboration with the project group. We searched for systematic reviews and meta-analyses, and replicated the search for primary studies conducted by Mir et al (10) for their systematic review in BMJ Open. The search for systematic reviews was limited to articles published in 2018. The complete search strategy, list of databases and websites and explanations are listed in Appendix 1 in the downloadable PDF.

Article selection and assessment of included studies

Two persons (AH and GH) independently reviewed all citations generated by the search for systematic reviews, to identify potentially relevant articles based on title and/or abstract. Full text versions were obtained for articles appearing to meet our inclusion criteria or for articles in which sufficient information was not available to make a decision. Two persons independently assessed the relevance of articles according to our list of inclusion criteria. Disagreements were resolved by discussion or by consulting a third party.

The methodological quality of systematic reviews meeting our pre-defined criteria was evaluated using the checklist for systematic reviews (11). All assessments were performed and agreed upon by two persons. Because more than one systematic review was rated as high quality, we chose between them on the basis of number of included studies/participants, and the date of their search for primary studies. The choice was also influenced by the clinical question formulated in the request for a national HTA.

Studies that attempted to disentangle the effects of antiplatelet therapy from the effects of anticoagulation therapy were preferred to studies that joined these two modes of treatment into a single composite comparator group. Network meta-analyses were preferred to meta-analyses containing only direct comparisons.

Data extraction

We extracted data as they were presented in the attached systematic review. When data were presented in several ways, we chose to report data in our preferred order: hazard ratio (HR), risk ratio (RR) and odds ratio (OR) with 95% confidence intervals (CI), or credible interval (CrI) in the case of network meta-analysis. When the included systematic review did not report data for our pre-specified outcomes, we retrieved the original publications to see if the outcomes were reported there.

Assessment of quality of evidence

We made use of the Grading of Recommendations Assessment, Development and Evaluation (GRADE) ratings presented in the attached systematic review. For all outcomes, the systematic review team assessed the certainty of evidence of benefits and harms of PFO closure compared to other treatments. GRADE allows a systematic and transparent critical appraisal of the potential limitations due to risk of bias, inconsistency, imprecision, indirectness and publication bias. We made use of Summary of Findings (SoF) tables from the selected systematic review and associated BMJ Rapid Recommendations, in an Infographic format and in MAGICapp (www.magicapp.org). The SoF-tables provide evidence summaries with relative and absolute effects across all outcomes and associated certainty of evidence.

GRADE gives the following definitions of the different quality of evidence:

  • High: Further research is very unlikely to change our confidence in the estimate of effect.
  • Moderate: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
  • Low: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
  • Very low: Any estimate of effect is very uncertain.

We presented the external experts with the GRADE SoF-tables and asked them to explicitly report disagreements with assessments made by the systematic review team. In the case of such disagreements, we planned to perform an independent critical appraisal of the identified studies and reported meta-analysis using the GRADE approach.

Health economic evaluation

We performed a model based cost-utility analysis (CUA) (12). The analysis was performed from a healthcare provider perspective, and discounted both costs and effects by an annual rate of 4%. We present main results as incremental cost-effectiveness ratios (ICERs), i.e. additional cost per additional unit of health. We measured costs in Norwegian kroner (NOK) and health effects in “years of good life” (quality-adjusted life years, QALYs).

Model structure

We designed a de novo health economic model in order to compare two alternative courses of action, interventional PFO closure and current treatment, with respect to health effects and costs. The decision analytic model can be viewed as two simulated cohorts (one for each treatment alternative) of patients that we follow for a given length of time, in this case until death. The model registers new ischemic strokes, major bleedings, deaths, functional status and health care utilisation for the patients in these two cohorts.

We set up the model as a Markov model with five health states. A specific health state represents a clinical situation that a person can experience for a shorter or a longer time. Health states were in this model defined based on functional status as measured by the modified Ranking Scale (mRS) , we included health states for mRS 0-2, mRS 3, mRS 4, mRS 5 and death (mRS 6). The mRS scale is described below (Table 1):

Table 1: Description of mRS health states (13)

We assigned each health state costs and health effects that accumulate as patients spend time in the specific state. Cost and health effects assigned to different mRS states are described below (Table 4). The model also included new clinical events, such as ischemic stroke, major gastro intestinal bleed and death. Similar to states, events can generate costs and health loss, but new events may also lead to transition to a different state. We illustrate model states, events and possible transitions in Figure 1.

Figure 1: Possible model health states (circles), events (squares) and transitions (arrows) for patients at different levels of disability

Individuals in this evaluation started in mRS state 0-2 (percentage of patients in mRS state 0-1 was approximately 81% and in RS state 2-3 approximately 19% in clinical trials (14)) and were propagated through the model based on transition probabilities estimated from epidemiological and clinical data.

Model parameters

Epidemiology

We based the risk of ischemic stroke on age specific rate of ischemic stroke from the Swedish stroke registry (15), to this rate we applied an increased risk of ischemic stroke connected to the PFO(16)and increased risk of reoccurrence (15)(Table 2). Risk of ischemic stroke caused by PFO was assumed to be higher among younger patients (age below 55) relative to older.

Table 2: Estimated rate of ischemic stroke for patients with PFO and a previous ischemic stroke event

For comparison, the estimated rate of ischemic stokes in the control arm of the Mir systematic review (10) was 0.02 for patients with an average age of 45.

Mortality hazards connected to health states are displayed in Table 4. Patients with light disability have a mortality risk similar to the normal population, while patients with severe disability have increased risk of dying. We applied hazards from Slot and colleagues (17) to tables from Statistics Norway. An increased risk of dying was also assigned to patients experiencing an ischemic stroke event or a major bleeding (15;18).

The percentage of patients below the age of 60 experiencing different levels of functional decline following an ischemic stroke were based on information from the Norwegian Stroke Registry (19), c.f. Table 3. Patients with an established functional impairment were assumed not to be able to improve after a new stoke, but would be able to remain at the same functional level or to experience a decline. Possible transitions for persons at different levels of disability are displayed in Figure 1. For example, a person with a functional level mRS 3, could either stay in this health state, or experience a new ischemic stroke. If this person experience a new ischemic stoke, this could lead to increased disability (transition to mRS 4 or transition to mRS 5) or the person could stay at the same level.

Table 3: Modified Ranking Scale (mRS) measured at 3 months after ischemic stroke for patients <65 years old. Numbers from The Norwegian Stroke registry 2015-2017 (19).

Clinical Efficacy

Current treatment for most patients is antiplatelet therapy. However, since anticoagulation may be considered a better treatment alternative for some patients; we also included this comparison.

In the health economic model, the driving efficacy estimates are risk of ischemic stroke and risk of major bleeding. These are the efficacy estimates generating the difference between the simulated cohorts receiving percutaneous PFO closure and usual care. With different numbers of individuals suffering an ischemic stroke or major bleedings in the two treatment arms, mortality and disability will also be different.

Clinical effect estimates were collected from the systematic review described in the clinical effectiveness section of this report. For the comparison with antiplatelet therapy, the estimates are for the events ischemic stroke and major bleeding, respectively OR=0.12 (95% CI 0.04 to 0.27) and OR=0.48 (95% CI 0.20 to 1.12). For the comparison with anticoagulation therapy, the estimates are respectively OR=0.44 (95% CI 0.08 to 3.83) and OR=0.26 (95% CI 0.07 to 0.82).

Costs

Costs in health states

Patients with functional decline will use different health and social care resources, e.g. admission to hospital, stays in rehabilitation facilities, home based rehabilitation and other (20).

Cost of ischemic stroke event and sequela health states were based on a Swedish cost study, that linked data from the Swedish stroke registry, Statistics Sweden, the National Board of health and Welfare and the Swedish Social Insurance Agency (21)(Table 4). This study includes data from 42,114 ischemic stroke patients, 17.9% of which were patients below 65 years old.

In order to fit with our purpose of analysingyounger patients (below 60) and applying a healthcare perspective as recommended in Norwegian Policy documents (22), information in Supplementary Tables S2 and S4 were combined to generate cost per mRS score for patients below 60 without including cost of work absence. Cost related to work absence constituted 2-34% of cost in the first year and 1-49% in the second year after ischemic stroke, with percentage of total cost clearly largest for the low disability groups (mRS 0-2). Disability costs measured at 12 months were allocated to states and generated as long as a patient spends time at this level of disability, while costs measures at 3 months are assigned as one-time transition costs.

Costs connected to health events

We based cost estimates related to the event major bleed from a previous evaluation of pharmacological anticoagulation treatment (23).

Cost of intervention and comparator

We estimated the total cost of PFO closure to be approximately NOK 113,000 (personal communication Elisabeth Leirgul and Lars Aaberge). Because the device price is confidential, we are not able to present costs disaggregatedly. Based on recommendations for pharmacological antiplatelet treatment and information from the Norwegian Medicines Agency, the annual cost of antiplatelet therapy (acetylsalicylic acid75 mg per day) is NOK 292 (24). The cost of anticoagulation therapy was based on prices of direct acting oral antagonists (DOACs). Drug cost of one-year treatment with DOAC is estimated to be approximately NOK 9,500.

Health Related Quality of Life

Quality of life multipliers for mRS health states were collected from Samsa et al. 1999 (25), values are displayed in Table 4. We applied multipliers (26) to population values as estimated by Burstøm and colleagues (27). Quality of life decrements connected to the events ischemic stoke and major bleed were based on previous work (23).

Table 4: mRS specific input

One way sensitivity analyses

In order to assess the robustness of the findings to changes in parameters, we performed one-way sensitivity analyses. We present the results as a tornado diagram, where parameters are ranked according to their impact results, with the most important parameter on top and the least influential parameter on the bottom.

Probabilistic sensitivity analysis

We assigned probability distributions to uncertain parameters following the approach described by Briggs and co-workers (30). We performed probabilistic sensitivity analysis based on Monte Carlo simulation by drawing random numbers from each probability distribution 10,000 times and recalculating the incremental cost-effectiveness ratio (ICER). We plotted the simulated ICERs on the cost-effectiveness plane and calculated probability of the interventions being cost-effective relative to comparator. Based on the same simulation, we also created cost-effectiveness acceptability curves illustrating the sensitivity of findings on cost-effectiveness of assumed equity adjusted estimates of alternative cost.

Organisational aspects

We evaluated the organisational consequences of a national introduction of PFO closure by consulting the clinical expert panel and relevant stakeholders.

Risks from radiation

The section about risks from exposure to radiation was written by the Norwegian Radiation and Nuclear Safety Authority.

Results

Clinical-effectiveness

Here we report the final evidence summary for both comparisons, followed by detailed results from the literature search, description of the included systematic review and relevant randomised trials as well as some detailed results from a network meta-analysis.

Tables 5 and 6 show the GRADE Summary of Findings for the two comparisons with relative and absolute effects, as well as the certainty in these estimates, across all patient-important outcomes. These SoF-tables are taken from the selected systematic review found to be most credible and informative for our health technology assessment. The external experts agreed on all GRADE assessments made by the systematic review team.

Table 5: PFO closure versus antiplatelet agents
Table 6: PFO closure versus anticoagulation agents

Results of literature search

We identified 18 titles in the systematic literature search for systematic reviews during the period January to August 2018. We reviewed the identified literature and found 13 references to be potentially relevant for our purpose, and full text copies were reviewed. Two systematic reviews with network meta-analyses met our inclusion critera (Figure 2), of which we selected the one with the most appropriate clinical question, design and methods as well as the highest number of participants and the most recent literature search. This selection process is in accordance with our established methods (11). For transparency, we note that the selected review was co-authored by one of the authors of the present report. List of excluded studies and reason for exclusion in Appendix.

Figure 2: Results of literature search

Description of the selected systematic review

The included systematic review by Mir et al. (10), investigated the effects of closure of patent foramen ovale (PFO) in patients with cryptogenic stroke, as compared to medical management with antiplatelet therapy or anticoagulation therapy. The authors conducted the systematic review within the context of a BMJ Rapid Recommendation, with a guideline panel defining which clinical questions to address and that assisted in the interpretation of the evidence(1).

The systematic review and network meta-analysis compared PFO closure + antiplatelet therapy with antiplatelet therapy alone, and with anti-coagulation therapy. We identified two separate comparisons to inform our health technology assessment, as presented below.

Their search for literature was executed on October 16th 2017, with one new study being added later. They aimed to include RCTs addressing the relative impact of PFO closure versus antiplatelet therapy versus anticoagulation in patients with cryptogenic stroke and patent foramen ovale. The authors included 10 RCTs from 8 studies, with a total of 4416 patients. We rated the systematic review to be of high quality, using the «checklist for systematic reviews» (11).

Table 7 provides brief descriptions of the included studies and their characteristics. Some of these studies used a composite comparator group where doctors assigned patients to either anticoagulation or antiplatelet therapy according to clinical judgement, whereas other trials assigned the participants randomly to one of the two options. The participants in the trials are predominantly less than 60 years of age, and a high proportion of participants had a large shunt size. The trials appear to have been conducted in relatively comparable populations, with no obvious causes of between-trial heterogeneity. Note that the CLOSURE 1 trial used the STARFLEX Septal Closure System, an earlier device that is no longer marketed. 

The multiprofessional team who conducted the systematic review concluded with moderate certainty evidence for key outcomes such as recurrent stroke, in particular given the risk of bias from lack of blinding as further detailed below. Their assessment was informed by the guideline panel who produced the BMJ Rapid Recommendations. This panel included experts, without significant conflict of interest but with sufficient clinical, methodological and research-expertise to perform appropriate critical appraisal of the body of evidence identified in the systematic review. The Norwegian external experts separately assessed the GRADE Summary of Findings from the systematic review and agreed on all judgments made concerning relative and absolute effects and the certainty of evidence for all patient-important outcomes.

Table 7: Characteristics of studies in selected Systematic Review (10)

Results from primary studies

The systematic review is based on data from direct comparisons in the included primary studies. We present these data as they appear in the original publications. We used the risk of bias evaluations performed by the authors of the systematic review. The authors noted risk of bias due to lack of blinding of medical personnel and patients regarding the placement of a PFO closure device. Also, half of the studies had incomplete data.

The results from the primary studies are shown in Table 8, and are presented as hazards ratios with 95% confidence interval when this was available. Otherwise, risk ratios with 95% confidence intervals were calculated by us derived from information contained in the tables in the original publications. No trials presented data on persistent atrial fibrillation separately from transient atrial fibrillation; these outcomes are therefore presented together as a composite outcome.

Table 8: Individual study results, reported as HR or RR (PFO closure vs medical management) with 95% confidence interval, or as the proportion of intervention group who experienced the adverse outcome.

Results from network meta-analysis

The authors in the included systematic review performed a Bayesian hierarchical fixed-effects network meta-analysis, with non-informative priors. Network meta-analysis combines data from different studies, using both direct and indirect evidence. For the comparison between antiplatelet therapy and PFO closure for the stroke outcome, the analysis was restricted to those trials where patients in the control arm were assigned randomly either to antiplatelet therapy or to anticoagulation therapy, or where they were assigned randomly to medical management and where at least 80% of patients in the control arm received antiplatelet therapy rather than anticoagulation. The PFO closure arm was chosen as the reference group. The report presents data for ischemic stroke, death, major bleeding, persistent atrial fibrillation or flutter, transient or paroxysmal atrial fibrillation or flutter, device or procedure-related adverse events, transient ischemic attack, pulmonary embolism and systemic embolism (See Table 9).

Table 9: Results from network meta-analyses, for PFO closure vs antiplatelet therapy vs anticoagulation therapy

Health economics evaluation

Model predictions of survival and ischemic strokes

If we assume no additional mortality connected to the ischemic stroke event or being in a sequela health state, the model predicts a life expectancy of 35 years for 45-year-old patients. When we account for the increased mortality connected to both ischemic stroke events and assign an increased mortality according to disability status as measured by the mRS scale, the predicted life expectancy drops to 27 years in the current treatment group (Figure 3). This means that compared to the general population, the population under evaluation will on average loose eight years of life due to their condition. Figure 3 illustrates the estimated survival curve for the simulated cohort of patients on current treatment.

Figure 3: Estimated survival curve for patients 45 years old with PFO and previous ischemic stroke receiving current treatment

If we simulate 300 patients 45 years old over a lifetime perspective, the PFO group is estimated to experience 125 new ischemic stokes, while the group treated according to current practice (medical management with antiplatelet treatment) will suffer estimated 261 strokes. This means that the health benefit of introducing interventional PFO closure in addition to antiplatelet therapy is expected to be 136 ischemic strokes prevented over a lifetime perspective per 300 patients treated. This result is a direct consequence of the estimated high efficacy of closure on new ischemic strokes (OR=0.12 95% CI from 0.04 to 0.27).

Estimated disease severity

When we calculate disease severity as absolute shortfall, the severity of disease is dependent on age. Patients, who are 45 year old, have a life expectancy of 30.6 “good life years” (i.e. QALYs) (31). With current treatment, patients 45 years old with a PFO and a previous ischemic stroke have a predicted prognosis of 15.8 QALYs, indicating a potential loss of 14.8 QALYs compared to the normal population. Following the approach described by Magnussen and co-workers (32), an equity adjusted f estimate of opportunity cost NOK 605,000 per QALY is suggested.

Incremental cost-effectiveness estimates

Compared to antiplatelet therapy alone, the addition of PFO closure results in a cost saving of NOK 755 339, and a substantial increase in health, 1.4947 QALYs (Table 10).

Table 10: Base case results for patients aged 45 years old, comparison with antiplatelet treatment. The table shows the absolute costs and QALYs over the lifetime of either alternative, as well as the incremental values and the ICER of adding PFO closure to antiplatelet therapy alone.

For some patients, the preferred treatment would be anticoagulation. Compared to anticoagulation therapy, PFO closure would generate 0.5958 QALYs, while generating a potential saving of NOK 637 195 (Table 11).

Table 11: Base case results for patients aged 45 years old, comparison with anticoagulation

One way sensitivity analyses

We display the result of the one-way sensitivity analyses in the form of a tornado diagram in Figure 4. As illustrated in the Figure, results are potentially sensitive to reasonable changes for a number of parameters. The most influential parameters are effect of PFO closure on risk of ischemic stroke, risk increase of stroke in the precence of a PFO, sec-ond year cost if spending time in the mRS 5 health state and risk of residive stroke.

However, the only change in single parameter that is likely to change the conclusion that PFO closure is a cost-effective alternative to antiplatelet therapy, is if PFO closure is ineffective in reducing the risk of ischemic stroke, this is the only change that will make the incremental net monetary benefits (INMB) cross over to a negative value.

Figure 4: Tornado diagram INMB PFO closure vs. antiplatelet therapy, illustrating the sensitivity of results for reasonable changes in single parameters. Results presented in incremental net monetary benefits (INMB), where all positive values indicate that PFO closure is cost-effective at the defined level of willingness to pay.

Probabilistic sensitivity analysis

Comparison with antiplatelet therapy

Result of the probabilistic sensitivity analysis is displayed in Figure 5. Each dot represents one possible combination from Monte Carlo simulation of incremental cost and effectiveness of PFO closure as compared to antiplatelet therapy. A visual inspection of the plot shows that PFO is likely to generate a larger health gain, while generating cost savings compared to antiplatelet therapy (exact numbers in Table 10).

In exact terms, PFO closure has a probability of 100 % of being more effective (i.e. generating more “good life years”) than antiplatelet therapy and a probability of 77% of being less costly. The estimated probability of PFO closure being a cost-effective alternative to treatment with antiplatelet therapy (assuming an equity-adjusted estimate of alternative cost of 605,000 NOK/QALY) is 98%.

Figure 5: Incremental cost-effectiveness scatter plot for PFO closure vs. antiplatelet therapy, patients 45 years old.
Table 12: Text report for Figure 5.
Cost-effectiveness acceptability curve

The cost-effectiveness acceptability curve (Figure 7) indicates that the conclusion that PFO closure is likely to be cost-effective compared to antiplatelet therapy is insensitive to the assumed equity-adjusted estimate of willingness to pay.

Figure 7: Cost effectiveness acceptability curve for PFO closure (triangles) vs. an-tiplatelet therapy (squares).
Comparison with anticoagulation therapy

Results of the probabilistic sensitivity analysis for the comparison of PFO closure vs anticoagulation therapy is shown in Figure 8. As we can see from the plot, there is more uncertainty as to whether or not PFO closure will be more effective, i.e. generate more “good life years”, than anticoagulation.

Figure 8: Incremental cost-effectiveness scatter plot for PFO closure vs. anticoagulation therapy, patients 45 years old.
Tabell 13: Text report for Figure 8.

In exact terms, our estimates indicate that PFO closure has a probability of 77% of being less costly than anticoagulation, and a probability of 79% of being more effective in terms of QALYs. Probability of being a cost-effective alternative (assuming an equity-adjusted estimate of alternative cost of 605,000 NOK/QALY) is 80%.

Cost-effectiveness acceptability curve

The cost-effectiveness acceptability curve (Figure 9) indicates that the conclusion that PFO closure is likely to be cost-effective compared to anticoagulation is insensitive to the assumed equity-adjusted estimate of willingness to pay.

Figure 9: Cost effectiveness acceptability curve for PFO closure (triangles) vs. anti-coagulation (squares).

Budget impact

If we assume that 300 patients (c.f. section under about organisational aspects) are eligible per year and that cost of PFO closure is NOK 113, 000, budget impact of implementation of the procedure would be NOK 33,900,000 per year. This estimate does not account for potential investments in increased capacity. Considering that approximately 135 of these 300 currently receive PFO closure, the marginal impact of implementation is NOK 18,645,000. The rate of new closures is assumed constant over time, impact per year will thus be the same.

Organisational aspects

Introduction of PFO closure at the national level may entail changes to existing routines, and may require investment in new facilities and new human capital. These consequences may affect both the general, stroke and cardiology units at local hospitals, and the university hospitals which will be conducting the intervention.

If a strong recommendation is issued for PFO closure in patients <60 years with stroke, most patients in this age range will be eligible for PFOclosure. Our deliberations on organisational consequences are built upon this scenario.

From the perspective of the local hospitals, the most important consideration arising from systematic introduction of PFO closure, will be the need to ensure adequate diagnosis (to identify patients with cryptogenic stroke and PFO) and referral for PFO closure. This requires that clinicians refer patients with cryptogenic stroke below 60 years of age to cardiological assessment with transesophageal echocardiography (TEE) to determine the presence and size of a patent foramen ovale and the extent of right-left shunt. Cardiologists then need expertise and capacity to perform TEE with satisfactory quality. Whereas we believe all hospitals, have cardiologists and equipment for TEE, the quality of the TEE procedures in the setting of PFO likely warrants further exploration. 

Further testing could be conducted at the local hospitals, including transcranial Doppler (TCD) with bubble test to determine the presence of right-to-left shunt. The diagnostic test accuracy of TCD, using TEE as the reference standard, was considered excellent in a meta-analysis of prospective studies (33). Another study found TCD study feasible in 91% of consecutive patients with TIA or stroke (34). It follows that this test could be introduced at the other university hospitals and local hospitals, pending on local expertise and access to equipment. We have not assessed organisational consequences of the introduction of TCD in more detail as it is not considered necessary, but rather a useful additional test to be considered introduced systematically in Norway, which is outside the scope of our report.

From the perspective of the hospitals at which the PFO closure is conducted, the organisational consequences may be appreciable. Today, this procedure is available only at Oslo University Hospital and Haukeland University Hospital in Bergen. Currently 135 closures are performed each year, but numbers seem to be increasing.

Norway experiences approximately 12,000 incident cases of stroke each year (2), of which 85% are ischemic strokes (35). Approximately 20% of strokes occur in patients aged 60 or younger, and 30% of these are estimated to be cryptogenic. In patients with cryptogenic stroke, with some estimates as high as 50%. With these numbers, one would expect that approximately 300 new patients would become eligible for this treatment each year. In order to meet this demand, increased capabilitites would be required.

There may be a «backlog» of patients who had a cryptogenic stroke before age 60, who may be eligible for PFO closure, but who were not referred at the time of their stroke. The age cut-off is set because PFO is more likely to be aetiologically relevant in young stroke patients; therefore, the key consideration is age at first stroke, not current age. For this reason, a large group of patients with earlier stroke may be eligible. This backlog may result in higher demand in the first few years after national introduction of the PFO procedure, however, it is unlikely that all of the eligible prevalent patients will be referred for closure. If PFO closure is going to be systematically offered to patients previously diagnosed with cryptogenic stroke and PFOclosurethe number of procedures will be substantially higher on a temporary basis.

Organisational aspects from a neurological perspective

At the time of this report, few stroke units in Norway utilize transcranial Doppler with bubble test to evaluate young patients (aged 15-60) with acute cerebral infarction. Traditionally, TEE has been considered the “gold standard”, and this method is crucial for verifying the presence of PFO. However, the diagnostics are improved when TEE is complemented by neurological evaluation including transcranial Doppler (33;36;37).

This requires ultrasound equipment, training of staff and accumulation of competence at the university hospitals and other stroke units that are involved in treating young patients with acute cerebral infarction.

Organisational aspects from a cardiological perspective

With a sensitivity of about 90% and specificity of 90% in studies (33), transesophageal echo (TEE) is considered the gold standard modality for diagnosing PFO. However, the examination is highly dependent on the skill of the operator.

Most local hospitals in Norway have available equipment for echocardiographic assessment including TEE, but the expertise of the cardiologists is varying with regard to examination of the atrial septum and PFO. The sensitivity might be improved by complementing investigations with transcranial Doppler bubble test. Nevertheless, training of cardiologists at local hospitals is necessary to ensure adequate assessment of the atrial septum and other cardiac structures before the decision on treatment. 

PFO closure has been performed by interventional cardiologists at OUS Rikshospitalet and Haukeland University Hospital. If new guidelines lead to a substantial increase in referral for PFO closure, it would be reasonable to offer this treatment in all the regional hospitals. This would require training of PFO teams (interventional and imaging cardiologists, and nurses) at St Olav´s University Hospital and possibly at the University Hospital of Northern Norway.

The estimated increase in PFO closure procedures from 125 annually today to 300 annually in Norway may have organisational consequenses. With close to 60 procedures pr million people, more university hospitals with heart surgical service in Norway will have sufficient patient volumes to offer PFO closure on a regular basis. These hospitals all have the necessary facilities for performing these procedures (cath labs, interventional cardiologists, TEE servce etc). However, this patient volume may have an impact on total capacity for patient treatment in cardiological departments as likely will demand more beds, lab capacity and staff. New centers will need training of staff to safely perform these new procedures.

Risks from radiation

PFO closure is a catheter-based transcutaneous procedure to prevent additional ischemic events after cryptogenic stroke. The procedure requires x-ray imaging and can potentially involve high radiation doses. It is therefore necessary to consider the risk related to exposure of both patient and operator. Radiation doses for patients and personnel depend on several factors such as equipment, working technique, experience and competence of personnel, use of protective equipment, complexity of the procedure and equipment for closure. The method for measuring PFO size will also be a relevant factor. The Radiation Protection Regulation (38) has requirements for all medical use of radiation. It is a prerequisite for implementing the method that the hospital meet these requirements.

Stochastic risk for patients

Dose statistics from the last 20 PFO closure procedures at Oslo University hospital and Haukeland university hospital have been collected and are presented in Table 14.

Table 14: Radiation exposure during PFO closure procedures

As the table shows, PFO closure results in less exposure than PCI given these data, and the radiation from PFO procedure at both hospitals can be considered to be low in dose. The two hospitals have relatively large difference in DAP-values, which indicate differences in equipment, technique or procedure optimization.

Risk for the patient

Considering the presented dose data there is little risk for deterministic effects. Regular radiation protection routines will be sufficient and detect any incidents involving over exposure. As for x-ray induced cancer, the increased risk of developing cancer compared to the population in general is negligible under these circumstances.

Risk for personnel

Personnel closest to the patient will be exposed to the highest radiation doses, especially cardiologists (REF). Organs most at risk are fingers and eye lenses (usually the left lens is exposed the most). Periodic monitoring of finger and eye lens doses are required. A personal dose meter attached to the left shoulder will give a good indication of the eye lens dose if no safety goggles are used.

It is important that adequate shielding equipment is used and that the personnel have competence to use it properly. Lead aprons and collar should be adapted to the current work situation and be personal. Use of safety goggles will significantly reduce the risk of induction of post capsular opacities and cataracts, if the goggles are ergonomically shaped and suitable for the cardiologist concerned. Lead curtains should be used on the side of the x-ray board, and ceiling mounted screens should be optimally positioned.

Personnel must also comply with regulatory dose limits, as proposed by The International Commission on Radiological Protection (ICRP) (41)and the Radiation protection regulations in Norway (38):

  • Skin / hands: 500 mSv (equivalent dose).
  • Eye lens: 20 mSv (equivalent dose).
  • Whole body dose: 20 mSv (effective dose).

Summary

Percutaneous PFO closure will normally give a moderate radiation dose to the patient and operator. The total dose (investigation, treatment and follow-up) for patients should be documented and evaluated, which is standard procedure in Norwegian hospitals performing such procedures. Any change in PFO closing technique and imaging procedures can affect the patient and personnel doses. This must be taken into account in the risk assessments prior to any relevant change in the procedures.

Discussion

Key findings summary

Key findings of systematic review

Patients diagnosed with cryptogenic (embolic) stroke due to PFO are at high risk of recurrent strokes (e.g. 10% over 5 years) being subsequently treated with currently recommended treatment (i.e. antiplatelet therapy). The selected systematic review included 6 randomized trials comparing PFO closure with medical treatment options, and 2 randomized trials comparing different medical treatment options with each other. In total, there were 3911 participants in these studies. In patients below 60 years PFO closure probably confers an important reduction in ischemic stroke recurrence compared with antiplatelet therapy alone (8%, moderate certainty evidence), but may make no difference compared with anticoagulation (low certainty evidence). PFO closure incurs a risk of persistent atrial fibrillation (2%) and device-related adverse events (3.6%). The most frequent device-related complications reported in the trials were vascular (1%), conduction abnormalities (1%), device dislocation (0.7%) and device thrombosis (0.5%). Compared with alternatives, anticoagulation probably increases major bleeding (2%).

Key findings of health economic evaluation

We find PFO closure to be a cost-effective alternative for secondary stroke prevention in patients with a previous cryptogenic ischemic stroke and diagnosed PFO. Health gains are most pronounced in the antiplatelet comparison, although gains in terms of QALYs are larger for both comparisons than what is usually observed in cost-effectiveness analyses (42). In both comparisons, the base case analysis indicates cost savings over a lifetime perspective.

Strengths and weaknesses

Possible limitations of systematic reviews

The selected review is based only upon randomized controlled trials, which is the optimal study design to inform questions about treatment effects. However the exclusion of non-randomized studies, such as large registry studies, may also result in failing to capture certain important clinical considerations, such as rare adverse events and complications in ordinary clinical practice.

Strengths and weaknesses of health economic evaluation

We have not included costs of informal care; although the effect on family members of functional decline in a loved one following an ischemic stroke may be significant (43). Inclusion of costs of informal care would have made the already very favorable results even stonger.

One simplification made in the health economic model is the assumption that patients are not able to improve their condition. This is a simplification of reality made in this modelling project; in reality, some persons may e.g. improve from mRS 3 to mRS 1. The earlier evaluation of thrombolysis included the possibility for 10% of patient to improve during the first year after a stroke, the 10% improvement probability was based on an expert opinion (44).

One weakness in this economic evaluation relates to the efficacy data on PFO closure as compared to anticoagulation. In the randomized trials, the chosen anticoagulation therapy was in 93% cases warfarin. The most used anticoagulation therapies for incident use is in Norway currently DOACs (45). The Norwegian Institute of Public Health has previously found DOACs to have a favorable profile compared to warfarin (23;46). If we believe that DOACs are more efficacious in preventing strokes while inducing less bleedings than warfarin, differences between PFO closure and DOACs would likely have been smaller, indicating more uncertain health economic results than found in the current analysis.

The cost of antiplatelet therapy is based on the assumption of monotherapy with ASA, a very inexpensive treatment alternative. According to national guidelines and current practice patients having undergone an ischemic stroke are, if they do not have atrial fibrillation, treated with ASA in combination with dipyridamole or with clopidogrel as monotherapy, rather than aspirin. As these are more expensive alternatives than ASA, the impact on the cost-effectiveness of PFO closure would be to further strengthen the already robust results. As illustrated in the tornado diagram (Figure 4), conclusions are very robust to changes in the cost of antiplatelet therapy.

Strengths of the analysis include strong registry-based input data for mortality rate, rate of ischemic strokes, costs and probability of different mRS states following an ischemic stroke.

Generalisability of findings

Overall completeness and applicability of evidence from the systematic review

The participants in the trials included in the systematic review were primarily younger than 60 years of age, and a large proportion of the participants had moderate or large PFOs or atrial aneurysms. We suggest caution against generalizing study findings to other groups of patients where the magnitude of the effect of PFO closure is expected to be lower (e.g. patients above 60 years of age, smaller PFOs, lower likelihood that the original stroke due to PFO). However, for patients matching inclusion criteria in trials the results should be generalisable, as reflected in the moderate certainty evidence where indirectness was not considered a problem in the GRADE evaluation (1;10).

Generalisability of findings from health economic evaluation

The health economic evaluation is specifically designed for the Norwegian context.

Results for comparison with anticoagulation may be different in a Norwegian context than indicated in this analysis, owing to the high uptake of direct-acting oral antagonists (DOACs) (47).

Consistency with other reviews

Consistency of systematic review with other reviews

At least 13 systematic reviews on the effects of PFO closure were published in 2018. These reviews all reach qualitative conclusions, which are broadly consistent with Mir et al. Due to differences in methodological approaches, other systematic reviews differ slightly in their estimates of the magnitude of the effect.

Consistency of health economic evaluation with other studies

Some health economic evaluations of PFO closure have been published in recent years (48-50) (Table 15). Similar to our analysis, these analyses find PFO closure to be a cost-effective alternative compared to medical management. Compared to our analysis, two of the analyses (48;50) differ in terms of finding increased incremental cost. One reasonable explanation for this divergence is the relatively high costs we have included in the mRS health states, making stroke prevention very favorable in terms of reduced cost. Cost estimates included in our analysis are from a very detailed study using real world data (21), this feature should be considered as a strength of this analysis. Variations in QALY gains estimated is within what is expected, considering differences in model structures and data sources.

Table 15: Published cost-effectiveness analyses

Implication of results on clinical practice

The national guidelines for stroke will likely be updated in 2019, based on the new and potentially practice-changing evidence for PFO closure. This process will allow clinical experts to weigh in on the implications of the new trials on PFO closure, for management of different groups of patients with cryptogenic stroke.

Need for further research

For the comparison of PFO closure vs anticoagulation the evidence only permits low certainty for the critical outcome of stroke recurrence, due to serious imprecision and indirectness. Further trials could lead to increased certainty and precision around the use of anticoagulation in particular, but also for clarifying the effectiveness of PFO closure in older patients and in patients with smaller PFOs. In addition to randomized trials on anticoagulation, follow-up studies, using observational data (e.g. large registry-based cohort studies) should further clarify the safety profile of PFO closure when performed in usual clinical practice, and add information on potential rare adverse events.

We have indicated some possible organisational consequences of national implementation of PFO closure. However, we would recommend that the regional hospital authorities carefully assess these challenges to ensure that patients across Norway receive equal access to high quality and safe PFO closure.

Conclusion

After combining data from eight trials on this topic, the systematic review by Mir et al, concluded that there is moderate certainty evidence for an important protective effect of PFO closure on risk of ischemic stroke, when compared to antiplatelet therapy alone. It is possible that some of this protective effect can be achieved with anticoagulation therapy. When using anticoagulation therapy as the comparator, the evidence for the effectiveness of PFO closure on risk of ischemic stroke only reaches low certainty. There is however, moderate evidence that PFO closure reduces risk of major bleedings compared to anticoagulation. PFO closure may based on this, be preferable to anticoagulation therapy for some patients.

Findings are likely to be specific to those groups of patients in whom the likelihood is high that the primary stroke was due to paradoxical embolism. This primarily includes younger stroke patients (<60 years). Further research is needed to clarify the extent to which other patient groups would benefit from PFO closure and to obtain higher certainty evidence on use of anticoagulation as an alternative to PFO closure.

We conclude that PFO closure is very likely to be a cost-effective alternative compared to medical management for stroke prevention in patients with cryptogenic ischemic stroke in a Norwegian setting. Conclusion is robust to changes in input data and consistent with findings in the published literature. PFO closure likely carries an acceptable risk from radiation. Introduction of PFO closure will have organisational consequences. Regional Health Authorities should further explore organisational aspects before implementation.

References

1. Kuijpers T, Spencer FA, Siemieniuk RAC, Vandvik PO, Otto CM, Lytvyn L, et al. Patent foramen ovale closure, antiplatelet therapy or anticoagulation therapy alone for management of cryptogenic stroke? A clinical practice guideline. BMJ (Online) 2018;362 (no pagination)(k2515).

2. Kvåle R, Forland G, Bakken IJ, Nguyen Trung T, Akerkar R, Dyngeland J, et al. Hjerte- og karregisteret: Rapport for 2012–2016 2018. Available from: https://www.fhi.no/publ/2018/hjerte--og-karregisteret-rapport-for-20122016/

3. Putaala J, Haapaniemi E, Metso AJ, Metso TM, Artto V, Kaste M, et al. Recurrent ischemic events in young adults after first-ever ischemic stroke. Annals of neurology 2010;68(5):661-71.

4. Rutten-Jacobs LC, Maaijwee NA, Arntz RM, Schoonderwaldt HC, Dorresteijn LD, van der Vlugt MJ, et al. Long-term risk of recurrent vascular events after young stroke: The FUTURE study. Annals of neurology 2013;74(4):592-601.

5. Waje-Andreassen U, Naess H, Thomassen L, Eide GE, Vedeler CA. Arterial events after ischemic stroke at a young age: a cross-sectional long-term follow-up of patients and controls in western Norway. Cerebrovasc Dis 2007;24(2-3):277-82.

6. Aarnio K, Haapaniemi E, Melkas S, Kaste M, Tatlisumak T, Putaala J. Long-term mortality after first-ever and recurrent stroke in young adults. Stroke 2014;45(9):2670-6.

7. Putaala J, Curtze S, Hiltunen S, Tolppanen H, Kaste M, Tatlisumak T. Causes of death and predictors of 5-year mortality in young adults after first-ever ischemic stroke: the Helsinki Young Stroke Registry. Stroke 2009;40(8):2698-703.

8. Helsedirektoratet. Nasjonal faglig retningslinje for behandling og rehabilitering ved hjerneslag [cited 22.12.2018]. Available from: https://helsedirektoratet.no/retningslinjer/hjerneslag/seksjon?Tittel=antitrombotisk-behandling-20014528#antitrombotisk-behandling-etter-hjerneinfarkt/tia-i-tilknytning-til-et-hjerteinfarktanbefaling

9. Verma SK, Tobis JM. Explantation of patent foramen ovale closure devices: a multicenter survey. JACC Cardiovasc Interv 2011;4(5):579-85.

10. Mir H, Siemieniuk RAC, Ge LC, Foroutan F, Fralick M, Syed T, et al. Patent foramen ovale closure, antiplatelet therapy or anticoagulation in patients with patent foramen ovale and cryptogenic stroke: A systematic review and network meta-analysis incorporating complementary external evidence. BMJ Open 2018;8 (7) (no pagination)(e023761).

11. Slik oppsummerer vi forskning: Folkehelseinstituttet [cited 21.01.2019]. Available from: https://www.fhi.no/kk/oppsummert-forskning-for-helsetjenesten/slik-oppsummerer-vi-forskning/

12. Petrou S, Gray A. Economic evaluation using decision analytical modelling: design, conduct, analysis, and reporting. BMJ (Clinical research ed) 2011;342:d1766.

13. Sacks D, Baxter B, Campbell BCV, Carpenter JS, Cognard C, Dippel D, et al. Multisociety Consensus Quality Improvement Revised Consensus Statement for Endovascular Therapy of Acute Ischemic Stroke: From the American Association of Neurological Surgeons (AANS), American Society of Neuroradiology (ASNR), Cardiovascular and Interventional Radiology Society of Europe (CIRSE), Canadian Interventional Radiology Association (CIRA), Congress of Neurological Surgeons (CNS), European Society of Minimally Invasive Neurological Therapy (ESMINT), European Society of Neuroradiology (ESNR), European Stroke Organization (ESO), Society for Cardiovascular Angiography and Interventions (SCAI), Society of Interventional Radiology (SIR), Society of NeuroInterventional Surgery (SNIS), and World Stroke Organization (WSO). Journal of vascular and interventional radiology : JVIR 2018;29(4):441-53.

14. Mas JL, Derumeaux G, Guillon B, Massardier E, Hosseini H, Mechtouff L, et al. Patent Foramen Ovale Closure or Anticoagulation vs. Antiplatelets after Stroke. N Engl J Med 2017;377(11):1011-21.

15. Korman M, Wisloff T. Modelling the cost-effectiveness of PCSK9 inhibitors vs. ezetimibe through LDL-C reductions in a Norwegian setting. European heart journal Cardiovascular pharmacotherapy 2018;4(1):15-22.

16. Alsheikh-Ali AA, Thaler DE, Kent DM. Patent foramen ovale in cryptogenic stroke: incidental or pathogenic? Stroke 2009;40(7):2349-55.

17. Slot KB, Berge E, Dorman P, Lewis S, Dennis M, Sandercock P. Impact of functional status at six months on long term survival in patients with ischemic stroke: prospective cohort studies. BMJ (Clinical research ed) 2008;336(7640):376-9.

18. Selak V, Kerr A, Poppe K, Wu B, Harwood M, Grey C, et al. Annual Risk of Major Bleeding Among Persons Without Cardiovascular Disease Not Receiving Antiplatelet Therapy. Jama 2018;319(24):2507-20.

19. Varmdal T, Norsk slagregister: evaluering av PFO lukking. to: Hagen G, 04.01.2019.

20. Fjaertoft H, Indredavik B. [Cost-estimates for stroke]. Tidsskrift for den Norske laegeforening : tidsskrift for praktisk medicin, ny raekke 2007;127(6):744-7.

21. Lekander I, Willers C, von Euler M, Lilja M, Sunnerhagen KS, Pessah-Rasmussen H, et al. Relationship between functional disability and costs one and two years post stroke. PloS one 2017;12(4):e0174861.

22. Meld. St. 34 (2015–2016). Verdier i pasientens helsetjeneste — Melding om prioritering Helse- og omsorgsdepartementet 2016. Available from: https://www.regjeringen.no/no/dokumenter/meld.-st.-34-20152016/id2502758/

23. Wisloff T, Hagen G, Klemp M. Economic evaluation of warfarin, dabigatran, rivaroxaban, and apixaban for stroke prevention in atrial fibrillation. PharmacoEconomics 2014;32(6):601-12.

24. Statens legemiddelverk[cited 2019]. Available from: https://www.legemiddelsok.no/sider/Legemiddelvisning.aspx?pakningId=6603f219-626a-40e1-8060-28d75ee4ec9a&searchquery=Albyl-E&f=Han;MtI;Vir;ATC;Var;Mar;Mid;Avr;gen;par;&pane=3

25. Samsa GP, Reutter RA, Parmigiani G, Ancukiewicz M, Abrahamse P, Lipscomb J, et al. Performing cost-effectiveness analysis by integrating randomized trial data with a comprehensive decision model: application to treatment of acute ischemic stroke. Journal of clinical epidemiology 1999;52(3):259-71.

26. Ara R, Brazier JE. Populating an economic model with health state utility values: moving toward better practice. Value in health : the journal of the International Society for Pharmacoeconomics and Outcomes Research 2010;13(5):509-18.

27. Burstrom K, Johannesson M, Diderichsen F. A comparison of individual and social time trade-off values for health states in the general population. Health policy (Amsterdam, Netherlands) 2006;76(3):359-70.

28. Valutakalkulator[cited]. Available from: https://www.norges-bank.no/Statistikk/Valutakurser/valuta/SEK

29. Priskalkulator[cited]. Available from: https://www.norges-bank.no/Statistikk/Priskalkulator/

30. Briggs A, Sculpher M, Claxton K. Decision Modelling for Health Economic Evaluation Oxford University Press; 2006.

31. Retningslinjer for dokumentasjonsgrunnlag for hurtig metodevurdering av legemidler Statens legemiddelverk2018.

32. På ramme alvor. Alvorlighet og prioritering. Helse- og omsorgsdepartementet 2015.

33. Mojadidi MK, Bogush N, Caceres JD, Msaouel P, Tobis JM. Diagnostic accuracy of transesophageal echocardiogram for the detection of patent foramen ovale: a meta-analysis. Echocardiography 2014;31(6):752-8.

34. Mazzucco S, Li L, Binney L, Rothwell PM. Prevalence of patent foramen ovale in cryptogenic transient ischemic attack and non-disabling stroke at older ages: a population-based study, systematic review, and meta-analysis. The Lancet Neurology 2018;17(7):609-17.

35. Fjærtoft H, Indredavik B, Mørch B, Phan A, Skogseth-Stephani R, Halle KK, et al. Årsrapport 2017. Norsk Hjerneslagregister 2017.

36. Mojadidi MK, Roberts SC, Winoker JS, Romero J, Goodman-Meza D, Gevorgyan R, et al. Accuracy of transcranial Doppler for the diagnosis of intracardiac right-to-left shunt: a bivariate meta-analysis of prospective studies. JACC Cardiovascular imaging 2014;7(3):236-50.

37. Pristipino C, Sievert H, D'Ascenzo F, Louis Mas J, Meier B, Scacciatella P, et al. European position paper on the management of patients with patent foramen ovale. General approach and left circulation thromboembolism. Eur Heart J 2018.

38. Forskrift om strålevern og bruk av stråling (strålevernforskriften), (2016).

39. Karambatsakidou A, Omar A, Chehrazi B, Ruck A, Scherp Nilsson J, Fransson A. SKIN DOSE, EFFECTIVE DOSE AND RELATED RISK IN TRANSCATHETER AORTIC VALVE IMPLANTATION (TAVI) PROCEDURES: IS THE CANCER RISK ACCEPTABLE FOR YOUNGER PATIENTS? Radiation protection dosimetry 2016;169(1-4):225-31.

40. Widmark A. Diagnostic reference level (DRL) in Norway 2017. Results, revision:and establishment of new DRL. . Norwegian Radiation Protection 2018.

41. The 2007 Recommendations of the International Commission on Radiological Protection. ICRP publication 103. Annals of the ICRP 2007;37(2-4):1-332.

42. Wisloff T, Hagen G, Hamidi V, Movik E, Klemp M, Olsen JA. Estimating QALY gains in applied studies: a review of cost-utility analyses published in 2010. PharmacoEconomics 2014;32(4):367-75.

43. Persson J, Levin LA, Holmegaard L, Redfors P, Svensson M, Jood K, et al. Long-term cost of spouses' informal support for dependent midlife stroke survivors. Brain and behavior 2017;7(6):e00716.

44. Stoinska-Schneider A, Robberstad B, Fure B. NIPH Systematic Reviews: Executive Summaries. In: Mechanical Thrombectomy for Acute Ischemic Stroke, Part 2 Health Economic Evaluation. Oslo, Norway: Knowledge Centre for the Health Services at The Norwegian Institute of Public Health (NIPH)

Copyright (c) 2016 by The Norwegian Institute of Public Health (NIPH). 2016.

45. Kjerpeseth LJ, Ellekjaer H, Selmer R, Ariansen I, Furu K, Skovlund E. Trends in use of warfarin and direct oral anticoagulants in atrial fibrillation in Norway, 2010 to 2015. European journal of clinical pharmacology 2017;73(11):1417-25.

46. Wisloff T, Ringerike T, Hagen G, Reikvam A, Klemp M. NIPH Systematic Reviews. In: Efficacy and Cost-Effectiveness of New Oral Anticoagulants Compared to Warfarin for the Prevention of Stroke in Patients with Atrial Fibrillation. Oslo, Norway: Knowledge Centre for the Health Services at The Norwegian Institute of Public Health (NIPH)

Copyright (c)2013 by The Norwegian Institute of Public Health (NIPH). 2013.

47. Kjerpeseth LJ, Ellekjaer H, Selmer R, Ariansen I, Furu K, Skovlund E. Risk factors for stroke and choice of oral anticoagulant in atrial fibrillation. European journal of clinical pharmacology 2018;74(12):1653-62.

48. Hildick-Smith D, Turner M, Shaw L, Nakum M, Hartaigh BO, Evans RM, et al. Evaluating the cost-effectiveness of percutaneous closure of a patent foramen ovale versus medical management in patients with a cryptogenic stroke: from the UK payer perspective. J Med Econ 2018:1-9.

49. Leppert MH, Poisson SN, Carroll JD, Thaler DE, Kim CH, Orjuela KD, et al. Cost-Effectiveness of Patent Foramen Ovale Closure Versus Medical Therapy for Secondary Stroke Prevention. Stroke 2018;49(6):1443-50.

50. Tirschwell DL, Turner M, Thaler D, Choulerton J, Marks D, Carroll J, et al. Cost-effectiveness of percutaneous patent foramen ovale closure as secondary stroke prevention. Journal of Medical Economics 2018;21(7):656-65.

51. Ando T, Holmes A, Pahuja M, Javed AA, Briasoulis A, Telila T, et al. Patent foramen ovale closure versus medical therapy to prevent recurrent stroke: A meta-analysis and trial sequential analysis. Journal of the American College of Cardiology Conference: 67th Annual Scientific Session of the American College of Cardiology and i2 Summit: Innovation in Intervention, ACC 2018;71(11 Supplement 1).