Scientific Advisory Committee on Medical Devices used in Cardiovascular Systems - Record of Proceedings - February 26, 2016

• Committee Members Present: John Ducas (Chair), Eric Cohen, Marino Labinaz, Alan Menkis, Barry Rubin, Raymond Yee, Joaquim Miró, Christopher Feindel
• Invited Guests: William Geerts, John Parker, Nancy Santesso, Jacqueline Saw
• Health Canada Representatives:
  • Office of Science: Hripsime Shahbazian, Caroline Hunt
  • Medical Devices Bureau: Kevin Day, Ian Aldous, Ian Glasgow, Jianming Hao, Karen Kennedy, Mark Korchinski, Christine Lefebvre, James McGarrity, Catherine Milley, Maurice Sylvain,  Lanyi Xu, Chris Schmidt
  • Therapeutic Products Directorate Director General Office:  Marion Law, Kimby Barton
  • Marketed Health Products Directorate:  Amanda Jones, Patrick Fandja
  • Health Products and Food Branch Inspectorate:  Barbara Harrison
• Regrets: Renzo Cecere, Brent Mitchell, John Webb, Toffy Mussivand

Abbreviations used in this record:

6-MHW:

6 Minute Hall Walk

AAA:

abdominal aortic aneurysm

ACP:

Amplatzer cardiac plug

AF:

atrial fibrillation

BAT:

Baroreceptor Activation Therapy

BMI:

body mass index

BVS:

Bioresorbable Vascular Scaffold

CHF:

Chronic Heart Failure

CRT:

cardiac resynchronization therapy

DAPT:

dual antiplatelet therapy

DES:

Drug Eluting Stent (DES)

DOAC:

direct oral anticoagulant

DVT:

Deep Vein Thrombosis

ICD:

implantable cardioverter-defibrillator

IFU:

instructions for use

IPG:

implantable pulse generator

IVC:

inferior vena cava

GRADE:

Grading of Recommendations, Assessment, Development and Evaluation

LAA:

left atrial appendage

LAAC:

Left atrial appendage closure

LVAD:

left ventricular assist device

LVEF:

left ventricular ejection fraction

LV:

left ventricle

MACCE:

Major Adverse Cardiac and Cerebral Events

MCUSA:

maximum circular unsupported surface area

MDB:

medical devices bureau

MI:

myocardial infarction

MS:

multiple sclerosis

MRI:

magnetic resonance imaging

NVAF:

non-valvular atrial fibrillation

OAC:

oral anticoagulant

OPC:

objective performance criteria

PG:

performance goal

PPM:

permanent pacemaker

RCT:

randomized clinical trial

RV:

right ventricle

SAC-MDUCS:

scientific advisory committee on medical devices used in the cardiovascular system

SAP:

special access program

SE:

systemic embolism

SVC:

superior vena cava

TAVI:

trans-catheter aortic valve implantation

TAVR:

trans-catheter aortic valve replacement

TEE:

trans-esophageal echocardiogram

TIA:

transient ischemic attack

TLF:

target lesion failure

TPD:

Therapeutic Products Directorate

VKA:

vitamin K antagonist

VTE:

venous thromboembolism

1. Opening Remarks & Welcome

Dr. John Ducas, Committee Chair
Marion Law, Director General, TPD

Dr. Ducas opened the meeting, welcomed the committee members and guest speakers. Marion Law, the current Director General of the Therapeutic Products Directorate (TPD), introduced herself, thanked the committee members and guest speakers for sharing their time and expertise and acknowledged the efforts all participants had made to attend the meeting. Ms. Law highlighted the importance of the Scientific Advisory Committee on Medical Devices Used in the Cardiovascular System (SAC-MDUCS) to Health Canada’s regulatory review and decision-making process.  She commended the committee on its continued support and noted that their past recommendations have been helpful in numerous license evaluations.  Ms. Law concluded by reviewing the topics of the day’s agenda and wished the committee members and guests fruitful deliberations.

2. Review of the Agenda, Affiliations and Interests Declarations, and Confidentiality Agreement

Dr. John Ducas, Committee Chair

The Chair reviewed the agenda items with the committee. The agenda was accepted with minor adjustments to accommodate all speakers. Members were asked to disclose any conflicts that may arise as the meeting proceeds.

Presentations are available upon request.

The responses outlined below are representative of the discussions/recommendations generated by the committee. 

3. Summary of how Health Canada has used information generated from previous SAC-MDUCS meetings

Kevin Day, Medical Devices Bureau (MDB)

Mr. Kevin Day presented a brief summary of how previous advice and recommendations provided by the committee have been considered by Health Canada.

The following topics were noted:

• Continuous Flow LVADs, thrombosis, anticoagulation, and pump speeds: committee feedback has been helpful in resolving post-market concerns.
• Thrombectomy Systems and the TOTAL trial: Recommendations from the committee have assisted the MDB in applying appropriate regulatory scrutiny in this area.
• LAA closure devices: the committee’s input and recommendations were taken into account for the licensing of the Watchman LAA Closure Device on 2016-01-19. 
• Human Factors and Cardiovascular Medical Devices: support from the committee on this topic has helped make this a basic requirement for a number of cardiovascular medical devices.
• Durability of Bioprosthetic Valves: committee feedback has been helpful in resolving post-market concerns and in assessing the need for longer term durability testing as the expected durability of bioprosthetic valves increases.
• Labelling considerations for IPG and significant risk/benefit factors:  Noted that the MDB is requiring more data to be included about the impact new features have on battery life, which is becoming a significant issue for emerging approaches for LV pacing with multiple vectors per beat.

It was noted that the Committee’s input ensures that good clinical oversight is continually being considered towards decision making processes for license, clinical trial, and special access applications.

Mr. Day thanked members for their ongoing dedication to the committee.

4. Leadless Pacemakers

Dr. Raymond Yee, Committee Member

Dr. Yee disclosed his affiliations and proceeded with his presentation.

He described a patient case study where the option of a leadless pacemaker was considered a lifesaving technology. The patient had abandoned leads, no superior venous access for new leads and had a previous epicardial system that failed. In this situation the leadless pacemaker was used as a rescue intervention providing an example of an ideal candidate for SAP.

Dr. Yee noted that there are many complications associated with traditional pacemaker leads, and leads are considered the Achilles Heel of pacing devices. The need to insert pacemaker leads for conventional systems can lead to complications. He indicated that these complications, especially in the acute phase, can be underreported, so accurate data on lead related complications are hard to obtain. Additionally, chronically implanted leads can fail structurally or fail to function effectively and need to be replaced while the non-functioning leads are usually left in the patient. These abandoned leads are always a contraindication for MRI scans and many implanted and functioning leads continue to be a contraindication for MRI scans. Lead failure is not limited to any one manufacturer and it is difficult to predict which lead models will have a higher failure rate.
 
Dr. Yee noted that based on published data (from the Danish registry where every lead and generator is tracked), a 9.5% complication rate at 6-months was reported. Lead-related complications were 2.8 -3% for new device implants or lead-related revision procedures.  However, even opening a pocket to replace a generator can stress and injure leads. In the Danish study, generator replacement had a lead-related complications rate of 0.9. The FOLLOWPACE Study reported a 5.5% lead-related complication rate in the first 6 months post-implant. Manufacturer-based databases are an added source of lead reliability but they often rely on physician reporting and therefore are of limited reliability. Nonetheless, the leads with the best long term reliability do become evident to physicians overtime. For instance the Sprint Quattro ICD lead is regarded as a very reliable ICD lead with an approximate survival rate of 95% at 10 years based upon data from the manufacturer.

Dr. Yee provided an overview of history of leadless pacemakers. He explained that the concept of leadless pacemaker is not new although sufficiently advanced devices ready for potential market release are only being tested.  Currently, several leadless pacemaker systems are available (although none are licensed in Canada), and are single chamber devices only. Dr. Yee described three systems that have undergone clinical testing: the LCP Nanostim by St. Jude Medical, the Micra by Medtronic and the WICS EBR.  He noted that both Nanostim and Micra devices offer single chamber RV pacing, with similar dimensions (weight of approximately 2 g and a volume of 1 cc). Both devices possess retrieval mechanisms. Dr. Yee indicated that pacing systems, either conventional leads or leadless devices, are placed wherever a stable location can be found with good electrical parameters in the patients’ myocardium. Implantation location is not limited to the apex, but should be a stable location with the best measurement of electrical parameters (sensing amplitude and pacing threshold/impedance).

On the other hand, the concept, design and purpose of the WICS leadless pacemaker system is entirely different from the Nanostim and Micra, and consists of two implanted components - an intracardiac device and subcutaneous device. WICS is designed to pace the LV in tandem with a separately implanted conventional right sided pacemaker or ICD. The intracardiac device is placed in the LV endocardium and does not have any self-contained energy source such as a battery power. Instead, the unit contains components that convert ultrasound energy to electricity. The subcutaneous portion of the WICS device system consists of a “generator” and subcutaneous pad “electrode” implanted in the chest wall that transmits ultrasound to the intracardiac hardware.  Dr. Yee described the design as ingenious, but complex, and noted that a recent study was terminated after 4 perforations and 1 death. Results of a second patient study have been reported but only involved 35 patients. There is scant other data available concerning the device at present.

Dr. Yee presented findings from the Nanostim Leadless II multi-center, prospective, non-randomized trial that involved 526 patients, with 6 months adverse events from the first 300 patients as the primary safety endpoint. He noted that most of the users had not implanted the device in humans before the trial. The study was limited to patients who did not require dual chamber pacing and were appropriate for single chamber (there is currently no ability to implant two leadless devices and communicate between the two).  The patient population in Leadless II was relatively elderly (mean age was 75).  He did note, however, that the BMIs were low, and, sadly not entirely reflective of a North American one.  He was impressed that there weren’t more perforations, given the age of the study group.

A question was asked whether there is a lower limit on patient size and if these devices could be used in a pediatric population. Dr. Yee estimated that leadless pacemakers could be implanted in patients as small as 20 kg but not less than 10 kg. This estimate was supported by members of the committee. Dr. Yee also expressed some concern about increased risks of perforations in the elderly with low BMIs, who are more fragile.

Dr. Yee stated that perforations are not uncommon, and that a major factor may be the delivery system rather than the device itself.  Stiff delivery sheaths make delivery difficult because of inferior handling and manipulability. He explained that the right ventricle is less than 2 mm thick in some locations and can therefore be easily penetrated. In future delivery system generations, performance could be improved with a decrease in size and increased handling of delivery sheaths.

The trial demonstrated a good implant success rate of 96% with 30% requiring repositioning after initial deployment. Dr. Yee commented that this statistic could be interpreted as evidence that the repositioning feature worked as intended.

Leadless II trial complication rates (526 patients) were as follows: 6.5% total, 1.5% perforation, 1.1% vascular complication, 1.5% dislodgement/migration, 0.8% retrieval/re-implant. Retrievals/re-implants were due to pacing threshold issues which occurred at 8±6.4 days.  Although the effectiveness threshold was set at 85% (85% or more patients would have acceptable pacing parameters at or below a given value) and met in this trial, a better level would likely be 90%.

The Leadless II trial primary endpoints surpassed performance goals for safety (freedom from serious adverse device effects (SADEs) at 6 months: 93.3% vs. PG of 86%), efficacy (acceptable pacing capture threshold (<= 2.0 V at 0.4 msec.) and a therapeutically acceptable sensing amplitude (R wave >= 5.0 mV, or a value equal to or greater than the value at implantation) through 6 months: 90.0% vs. PG of 85%), and procedure/implantation (successful device implantation: 93.4% vs. PG of 85%).  In addition the investigators reported decreased capture thresholds between implant and pre-discharge, and out to 12 months, which Dr. Yee interpreted as a good sign, instead of increasing over time as is often seen with conventional leads. Also, impedance significantly decreased from implantation to 12 months. While Dr. Yee considered these findings supportive of safety and effectiveness, he stated that higher impedance (R) values would provide better battery longevity.

There was a comment that a perforation rate of 1.6% seems high compared to that of single right ventricular leads and further details for cardiac perforation and clinical implications were asked. Dr. Yee explained that perforation with conventional leads is probably under-detected and under-reported. This complication is seen most often in the elderly and when it occurs, the implanter simply withdraws the lead and repositions: rarely does it result in clinically significant sequelae such as cardiac tamponade or prompt further investigations to document if the patient is hemodynamically stable and asymptomatic. Even when it is significant, these events are not reported except in the context of a clinical trial or mandatory registry.

There was a question whether there would be an interaction between these devices and previous devices already implanted in the chamber, such as previously placed leads, which may create electrical noise and impact sensing. Dr. Yee commented that contact between hardware can generate electrical noise and potentially interfere with pacemaker function. The solution is to place devices far enough apart and there is sufficient space in the RV for multiple devices; he also thought that the size is such that it would be feasible to implant 3-4 devices in an individual within their lifetime.

Dr. Yee described Micra leadless pacemaker, and indicated that while it is sized similarly to the Nanostim, it has a no magnet mode. This is a drawback as pacing cannot be forced in an emergency.

The Micra trial (Reynolds et al. 2016) enrolled 725 patients with characteristics similar to those included in the Leadless II trial, had a very good survival curve with 4% major complications at 6 months which Dr. Yee described as acceptable.  Micra trial complications included: 1.6% perforation, 1% vascular, 0 dislodgement/migration, 0 retrieval/re-implant, and one death attributed to metabolic acidosis. Dr. Yee felt that the large sheaths in elderly patients could be a major factor in the perforations. The Micra trial showed a good pacing threshold with a time course that was remarkably stable.

It was noted that battery life on leadless pacemakers is approximately 10 years based on pacing voltage of 1.5 V. Dr. Yee indicated that higher impedance spares battery life because of less current loss (Ohm’s Law). The electrical performance data of the Micra study are remarkably similar to the Nanostim study.

There was a question on need for steroid coating similar to leads. It was noted that both systems include a steroid coating. Dr. Yee commented that the value of steroids is in blurring the spike in pacing threshold that occurs usually within the first three months with non-steroid coated leads. The time profile seen so far in Nanostim and Micra may partially reflect the influence of steroid coating but, so far, no signal of an acute rise in pacing thresholds post-implant has been observed. More data will be required before the value of steroid coating can be re-evaluated in the context of leadless pacemakers.

A discussion on the fixation of leadless devices followed. There was a comment that without good fixation movement is of potential concern due to interaction with structures such as chordae. Dr. Yee commented that these were early days but did seem that both Nanostim and Micra fixation mechanisms seem to be working well with a low dislodgement rate. The Nanostim fixation is a screw similar to leads. The Micra may experience less motion due to the 4 fixation tines. It was asked whether use of imaging would be encouraged to see if there is movement or interaction with other devices. Dr. Yee noted that interaction of pacing leads with chordae and valve leaflets is rarely noted but it is also true that it is rarely looked for with conventional leads.  For Micra implants magnified fluoroscopic imaging is looked at during traction on the tether to test for fixation but interaction with any right heart structure might be minimized by the presence of the delivery sheath and would not be seen until it was removed. Moreover the leadless pacemaker implants typically target sites deep in the RV such as the junction of the septal and free walls so such interactions would be unlikely.

Retrievability/extractability of these devices was discussed next. Dr. Yee noted that extraction of a leadless pacemaker is easier than traditional leads. In the Leadless II trial, 7 retrievals were successful, although these were short term implant retrievals. Devices implanted for 5 years may be much harder to explant but are still thought to be easier and less risky than conventional leads. Conventional leads are usually fibrosed and adhered to the vein wall along their lead bodies and laser lead extraction can lacerate the central venous structures in addition to tearing the RV myocardial wall. It was also suggested that the need for extraction of the leadless devices may be less, due to less infection than experienced with leads.  Devices will not likely be removed if there is no problem. It is possible to implant multiple leadless pacemakers side by side within the same chamber.  Dr. Yee noted it is feasible to select points that will pace and that 3-4 devices would fit in the chamber.

There was a concern regarding the severity should leadless devices detach; the worst case scenario being pulmonary embolism.  Dr. Yee commented that this is a legitimate concern and there is a need to see minimum 12 month data on dislodgement to determine if fixation is secure.  He was of the opinion that if it hadn’t dislodged by 12 months than it is unlikely to ever dislodge.

A question was asked if thrombosis is observed on the leadless devices and whether there is a need for anticoagulation. Dr. Yee commented that thrombus is frequently observed on conventional pacing leads and they come and go, as thrombi are continually formed and absorbed. It is still an open question whether they are clinically important (no evidence of poorer outcomes) and whether anticoagulation is required.

He summarized his presentation noting that:

• Device technology has come a long way.
  • Leadless cardiac pacing is feasible in humans and evolving rapidly
  • Shows good short-term efficacy;
  • Has similar short term safety profile to transvenous pacing;
• May provide an alternative for those with difficult transvenous or epicardial access
• Long term data still required
  • Safety and device performance
  • Extractability

At the end of the presentation Dr. Yee addressed the questions provided by Health Canada.

Question 1: Who is the target patient population for these devices? Will these devices completely replace single-chamber pacemakers?

Dr. Yee talked in terms of near term and long-term target populations. In the near term, he considered these devices feasible for select patients now because they have demonstrated good pacing performance with reasonable procedural safety. They are reasonable as a bail-out option since such patients have little alternative. Although initially considered for those with limited venous access, high infection risk, and one device lifetime, he specifically noted that it is a reasonable option for patients being considered for epicardial or abdominal PGs since these devices are fraught with problems and often last less than 5 years. Leadless pacemakers are also an option for those who require a second RV device since there is room for multiple devices.  Nevertheless, there is a need to be cautious about patients targeted and build experience going forward. He did not think these devices would replace existing pacing systems in the short-term, until more longer term performance, safety and extractability data show equivalent safety and performance to conventional pacemakers.

There was a question about the active patient and rate response. Dr. Yee commented that there is not a lot of data on this and we need clinical data to demonstrate that leadless pacemakers are as good as conventional pacemakers with respect to rate response (i.e. physically active patients).  Dr. Yee noted that the Micra device tries to filter out the heart motion, whereas the Nanostim bases its algorithm on venous blood temperature.  Dr. Yee suggested that manufacturers of these devices will need to demonstrate their performance with a standardized test (e.g., treadmill test) in response to exercise and physical activity.

Question 2: The useful life of the device is probably in the range of ten years. Tissue growth may prevent explantation. Is this a problem?

This was answered throughout presentation.  Extractability is certainly an issue but there may be trade off in terms of less risk.  The devices will not require extraction if there is no problem like infection.

Endothelialization is expected with any intravascular hardware, leadless or leaded. The dimension of the issue will change but the risks of extraction may be the same or less because:

• No lead to fibrose to SVC wall or thrombose SVC
• Extraction sheaths will be similar to current tools but larger until device size decreases further.

Question 3: Are there specific contraindications, warnings or precautions that should appear in the labeling.

• Significant risk of cardiac perforation and highlight embolization risk.
• Restricted to use in patients who cannot receive standard PPM
• Not for use in RA or LV.
• Use with caution in patients with TV prosthesis; not for use with TV mechanical prosthesis

Question 4: Will there be a learning curve? What training is required? Who should implant them?

Yes, there is a steep learning curve related more to the maneuverability of the delivery sheath. It is much different than a lead implant and operators need experience with sheaths and catheters.  It was recommended to have initial animal or simulator training, followed by proctoring for the first 4 to 6 cases.

Question 5: Can you predict who will need dual (or more) chamber pacing or ICDs. If you could predict, which type of system would you prefer to implant?

The indications for pacemakers and ICDs are fairly clear and codified in practice guidelines. Put simply, if the LVEF < 30-35%, an ICD is indicated, not a leadless pacemaker. A CRT device (CRT-P or CRT-D) is indicated in heart failure patients with EF<40% with a wide ORS.  The use of dual chamber pacemakers is still controversial. Randomized comparative trials have failed to show clinical benefit from dual chamber pacemakers compared to single chamber pacing with the exception of lower rates of AF. Despite this, many physicians still believe there exists an important role for dual chamber pacemakers and clinical practice varies by center and implanter.

Question 6: How should regulatory requirements for safety and effectiveness be resolved for this new technology, e.g. a randomized clinical trial, registry etc. Describe potential controls and trial safety and effectiveness endpoints.

A RCT would best support safety and effectiveness in an ideal world. That being said, safety can be demonstrated without a RCT using good historical controls and conservative estimates for conventional pacing systems. If the devices are indicated for a bailout indication (e.g. patients who are ineligible for standard pacemaker), then RCT not likely required.  If leadless devices are considered for replacing standard systems, something closer to a RCT may be required.

The following endpoints were indicated:
Safety - acute procedural complications: perforation, vascular access (pseudoaneurysms/aneurysms), major bleeding or hematoma, early dislodgments, embolization
Efficacy - successful deployment of LLPM, acute and chronic electrical performance, need for re-operation at 12-24 months.

Question 7: Are there some OPCs that you could recommend for performance or should Health Canada require RCTs to study this new technology.

No standard OPC exist, however the following is recommended and should be an amalgam of criteria rather than focus solely on one safety or efficacy endpoint (Kirkfeldt Eur HJ 2014):

• Need for re-intervention @ 6 months <2.5%
• Dislodgment (<2.5%), cardiac perforation (<0.6%)
• High pacing threshold requiring reoperation (<1%) OR
• Battery longevity is the critical clinical endpoint
• >85%pts will have pacing threshold (_V@_ms) at 12 months that allows for battery longevity ≥7 years (100% VP, impedance of 500 ohms and programmed pacing output safety margin of ≥2)

Battery longevity should be considered as a performance criterion under real world conditions and not just the idealized labelled value with a given pacing threshold of X.  Bench testing according to ISO standard tests is not enough, the manufacturer should demonstrate that the real world longevity is reachable.

Dr. Miro joined the meeting.

5.  Patient Preference

Dr. Nancy Santesso, Guest Speaker

Dr. Santesso introduced herself and disclosed her affiliations.  She noted that she is a member of the GRADE (Grading of Recommendations, Assessment, Development and Evaluation) working group which uses the GRADE approach, a systematic method, of creating guidelines.

She was asked to address patient preference in informed health decision making:

In the past few years, there has been increased interest focused on the importance of allowing for better patient preference in informed health decision making. Health Canada is starting to explore how we might incorporate patient preference considerations during premarket review. Patient preference may be defined as: qualitative or quantitative assessments of the relative desirability or acceptability of attributes that differ among alternative diagnostic or therapeutic strategies.

She began her presentation indicating that Health Canada is already accounting for patient preferences; however, could be more formal, could use better data, could be more transparent, and could learn from others in Canada.

An example was given using a lay scenario to describe how patient preferences are accounted for. The GRADE method was described.

• The important patient outcomes need to be decided at the beginning. Only these outcomes are considered in the decision.
• Values are added to each outcome and there is a balancing act to determine which outcomes hold more weight; consideration must be given to whose values will be used.
• Outcomes are decided using input from our own values, values of other people present, and from asking others.
• If someone identifies another outcome later in the process it is not considered in the decision since it was not identified as an important outcome.

A question was asked about the reliability of data from surveys. Dr. Santesso described a hierarchy for obtaining documentation and referenced the FDA guidance document, which she described as quite well done.

A question was asked regarding how design changes over the lifecycle of a product are accounted for in the decision making if all important outcomes must be identified at the beginning. Dr. Santesso responded that feedback for new critical outcomes can be factored in if new data reveals their need.  This requires going back and making the decision over again. There needs to be transparency on the weight of each outcome.

Dr. Santesso presented on how to determine the critical and important outcomes for the decision. This was focussed on methods including scoping literature and surveys. Potentials for surveys include patient groups, professional groups, and program managers.

Patient values are used to weight the outcomes. Patient values can be obtained from:

• Asking the panel to reflect on experiences making decision with patients
• A patient representative as a panelist
• Consulting patient representatives
• Performing a value rating exercise of patients
• Original studies of rating values of patient
• A literature review of those studies

The first three on the list are similar to focus groups or unsystematic observational studies.  Patient values can also be elicited from studies and qualitative research.

Dr. Santesso noted that separate panels with patients and citizens may be necessary if having a patient at the table may change the conversation.

Dr. Santesso highlighted that it can often be difficult to weigh the patient population choice (obtained via panels and surveys) versus the individual patient choice. Patient preference is one factor of a decision.  There is a need for transparency on how it is considered.

SAC members noted that often clinicians can convince patients of just about anything.

She suggested list of key questions to consider when making recommendations:

• Is there important uncertainty about or variability in how much people value the main outcomes? This should be clear to potential users - clinicians and patients. If so, these outcomes must be addressed on an individual basis. If not, a panel can address the outcomes.
• Is the intervention acceptable to key stakeholders? This should be clear to potential users - clinicians and patients.

The overall risks and benefits become desirable and undesirable consequences. This is a balancing of benefits X value, acceptability, feasibility versus risks X value, resources, equity.

1. Should Health Canada account for patient preference? We are used to thinking of interventions and devices in terms of Risk and Benefit. How could we capture patient preference as a “benefit” and how can we assess this against a possible increased risk. Note: Health Canada would still have to be confident in the safety and effectiveness or appropriateness of a device prior to issuing a licence or authorizing special access or a clinical trial.

• Risks and benefits can be weighed using patient values (or other key stakeholders)
• Patient preference as an outcome of factor in decision making

2. Would it be helpful if the use of patient preference data was included in the labelling, or elsewhere (for example, Summary Basis of Decision) to identify this as a factor in the regulatory decision.

• Make outcomes and criteria transparent to potential users of device.  This includes a clear statement on what was considered, including values and preferences.
• Useful in individual decision making
  • Clinicians should recognize that different choices will be appropriate for each individual and that clinicians must help each individual arrive at a management decision consistent with his or her values and preferences.

3. How is patient preference accounted for in the Canadian clinical environment? Is it generally included in Canadian Clinical Practice guidelines? How would an assessment of patient preference at pre-market review fit with current practice?

• GRADE approach to clinical practice guidelines
• Institute of Medicine Standards

General discussion on the topic of patient preference ensued.  Patient preference and its relation to informed consent were discussed. It was noted that the main concern here is whether the values of the patient are the same as the physician.  There was a question whether it would be beneficial for Health Canada to require informed consent similar to what is done in clinical trials.  General consensus was that this is already done and physicians have a legal obligation to disclose risks to the patient. It was noted that some of this is regulated at the provincial level already.  It was further commented that there is no ability to monitor the consent process.

Discussion evolved to how physicians can be better prepared to inform patients and where this information can be found.

• There is a need to convert clinical trial data into more meaningful information for patients to understand how they may personally benefit.  For example, replace “numbers needed to treat” with “this will increase your life expectancy by x”
• Physicians could use guidance on factors to discuss with a patient that they might not think to consider.
• Anything to enhance the process of a physician having information to give to a patient is of benefit.
• This information would be “hidden” in the labelling - not a useful source.
• Medical associations are a good source of information.
• May be of value to more closely monitor these types of outcomes in clinical trials
• An unbiased patient pamphlet that does not come from the manufacturer could be helpful. Medical associations perhaps are a good place to develop such information brochures.

6. Considerations with LAA Occluders

Dr. Jacqueline Saw, Guest Speaker

Dr. Saw introduced herself and disclosed her affiliations.

Dr. Saw was asked to provide a review of the use of Left Atrial Appendage (LAA) occluders including a discussion about the appropriate anticoagulation/antiplatelet regimen prior to and following implantation.

She started by providing an outline of her presentation:

• Background rationale for LAA closure
• Landscape of LAA closure
• Devices available: endovascular and epicardial
• Scientific evidence for endovascular LAA closure: OAC eligible and ineligible patients
• Cost-effectiveness analysis of LAA closure
• Discussion on proposed questions

Background Overview of Why LAA Closure is performed

Dr. Saw explained that LAA closure is being positioned as a possible alternative to oral anticoagulant (OAC) therapy for atrial fibrillation. Functional changes, structural remodeling, blood stasis and potentially hypercoagulability can occur within the left atrial appendage (LAA), especially during atrial fibrillation (AF), and this can result in thrombus formation.  Thrombi formation in the LAA has been identified in over 90% of patients with nonvalvular atrial fibrillation (NVAF) in various studies, mostly by using trans-esophageal echocardiogram (TEE).  Particular anatomical and functional characteristics of the LAA in adults contribute to the occurrence of thrombus. LAA morphology between individuals is highly variable, but may be described as long, tubular, or multi-lobed; it is a blind-ended pouch with excessive trabeculations which can cause blood stasis. Dr. Saw indicated that different anatomies would require different devices.

She also indicated that thromboembolic events, because of AF, lead to stroke, and that 35% of patients with AF will suffer a stroke in their life time. She also suggested these events would result in relatively high rates of disabling stroke or mortality.

Clinical guidelines recommend that most patients should be treated with anticoagulation as the standard of care, unless they are at low risk or have a contraindication to anticoagulation.  Dr. Saw presented data from the Registry of the Canadian Stroke Network (2003-2007) that showed 90% of AF patients who had a stroke were not receiving Guideline-recommended treatment using Oral Anticoagulants (OACs), and in that population, 30% received suboptimal anticoagulation, 30% received antiplatelet therapy, and 30% received no treatment.
 
Meta-analysis studies have shown that New Direct OACs (DOACs) have demonstrated a reduction in the occurrence of stroke or systemic embolization compared to Warfarin. The real clinical benefits, however, have been limited because:

• ~30% eligible patients are not on OAC.
• OACs can be contraindicated, not tolerated and subsequently discontinued in 20% to 40% of patients,
• There are bleeding risks
• Long-term adherence is suboptimal

Landscape of LAA Closure

It is projected that there will continue to be an increase in the population affected by AF, which is the most common adult arrhythmia.  In 2000, there were 5.1 million in the United States affected by AF; by 2050 projections indicate there will be 15.9 million with AF.  This represents 1% to 2% of the total population. As such, Dr. Saw indicated the LAA closure market is expected to grow 6x.

LAA Closure Devices Available

Dr. Saw noted that currently, there are a number of endocardial (percutaneous) and epicardial LAA closure devices available on the market or in the development stage.

The first generation of endocardial closure device PLAATO has been removed from the market while the 2nd generation of closure devices are available in European market including Watchman (and Watchman FLX), Amplatzer Cardiac Plug (ACP), Amulet, and WaveCrest. Several other endocardial LAA closure devices are being tested in the preclinical trial evaluation.

Two epicardial LAA closure devices, Lariat and AtriClip, have been approved by FDA and CE marked, and two other devices are either in the clinical trial evaluation or pre-clinical trial evaluation stages.

Watchman was CE marked in 2005, approved by FDA in March 2015, licenced by Canada in January 2016, and is now available in over 55 countries with 15,000 implants. Watchman is a nitinol frame with 5 sizes ranging from 21 to 33 mm (2-4 mm oversize). It is delivered through transseptal puncture via 14 Fr sheaths, and is repositionable and retrievable. It can be expanded radially to maintain its position within the LAA. This device is covered with a 160 µm PET membrane and promotes healing (by endothelialisation) within 45 days.

Watchman FLX has been designed in order to simplify the procedure and minimize adverse events. Specifically, Watchman FLX has a shorter device length than the Watchman with decreased recapture forces and better conformability.  It received CE Mark in Nov 2015.
 
ACP was CE marked in 2008, is used in Canada under the Special Access Program (SAP), with approximately 10,000 implants worldwide.  ACP is constructed of a nitinol mesh and Dacron polyester sewn into both the lobe and disc, connected by a central waist.  It has 8 sizes 16-30 mm (3-5 mm oversize). The shape fits most LAA, and it is retrievable and re-deployable, until it is released.

Amulet (2nd Gen ACP) has a recessed end screw, a larger disc diameter, a longer lobe length, a longer waist length, and sizes up to 34 mm.

Other LAA closure devices including WaveCrest, Ultrasept, PFM Device, and LARIAT were described briefly.  It was noted that the LARIAT is generally not used in Canada.

Scientific evidence for endovascular LAA closure: OAC eligible & ineligible patients

Dr. Saw noted that clinical evidence for the safety and effectiveness of endovascular LAA closure devices has been collected via an RCT on the stroke prevention in patients treated with endovascular closure devices using WATCHMAN compared with warfarin.

WATCHMAN clinical trial program includes 8 completed or ongoing trials with a total of over 3700 pts with over 6000 pt-years follow-up data currently available.

• Pilot
• PROTECT-AF Study (N=707)  
• CAP Registry
• ASAP
• PREVAIL
• CAP2 Registry
• EWOLUTION
• WASP

It was noted that in the PROTECT-AF Study (Reddy et al. 2014) that results at 3.8 years were promising.  Although some acute events were higher than anticipated (e.g. 4.8% pericardial effusion), these were partially attributed to the learning curve.

There was a question about the interpretation of the PREVAIL RCT (Holmes et al. 2014) data where 2 of the 3 endpoints were not met.  Dr. Saw indicated that the data was supportive of safety and effectiveness of the Watchman when taken in the totality of the evidence.  In the PREVAIL RCT data, outcomes in the warfarin arm, which was relatively small, reflected a very low stroke rate compared to the expected outcomes seen historically in this patient population.

Real World Experience

A prospective real-world registry study EWOLUTION registry (Boersma et al. 2016) included 1021 patients in 47 centers in Europe, Russia, Middle East with follow-up for 3 months and included 61.8% ineligible for OAC (38.7% prior major bleed or predisposition to bleed). DAPT was used in 59.5% post-implant (OAC 27%). The Procedural Success was 98.5%. There were 2.7% serious procedural/device-related safety events to 7 days including perforation, tamponade, ischemic stroke, device embolization, and vascular complications. The 30-day mortality was 0.8%.

Canadian WATCHMAN Experience

A Canadian Multi-center Experience with Watchman for percutaneous LAA closure presented by Dr. Saw at AHA 2015 included 106 patients with atrial fibrillation.  At baseline, only 32% of patients were on any anticoagulant, with 13% on no antithrombotic medication and 39% on a single antiplatelet medication.  Procedural Success was 97.2% and the Composite Major Safety Events at 7-days was 0.9% (Composite of all-cause death, ischemic stroke, SE, or device-/procedural-related events requiring surgery or major endovascular intervention within 7 days). The rates of stroke rate, TIA, and systemic embolization (SE) were 0%, 1.9%, and 0%, respectively. At discharge, there were 73.4% on DAPT alone. The Canadian pooled stroke/TIA/SE rate of 3.3 %/Yr was reduced by 66% and 59% when compared to the estimated rates by CHADS-VASc and CHADS, respectively.

German Study Data

The long-term results of Watchman LAA closure with some data up to 5 years (mean follow-up period of 3.0+- 1.6 years) showed the rates on Post-procedural DAPT for 6 months and Post-procedural VKA and Aspirin were 41.8% and 58.2%, respectively. The annual rates of TIA, stroke, intracranial hemorrhage, and death were 0.7%, 0.7%, 1.1%, and 3.5%, respectively. The study concluded that LAA closure with the Watchman device is safe and feasible for stroke protection in patients with atrial fibrillation. Low ischemic events rates demonstrate its effectiveness during long-term follow-up (Wiebe et al. 2015).

ACP Registries

There have been a number of registries on ACP worldwide with a total of 1058 patients included, including a registry of 52 patients in Canada. The averaged technical success has been 97.2%. The annual stroke is 2.3%, which is reduced by 59% based on estimated CHA₂DS₂-VASc score. The annual major bleeding is 2.1%, which is reduced by 61% based on HAS-BLED score. The rates of embolization and thrombus formation were 1% and 4.4%, respectively.

A multi-nation, multi center AMPLATZER Amulet observational post-market study with proposed 1000 subjects is ongoing.  Enrollment may be completed in 3 months. 
 
The results from real-world LAA studies on 976 subjects treated with DAPT for various durations ranging from 1 month to 6 months were also presented.

Cost-Effectiveness of LAA Closure (LAAC)

A general overview of the cost effectiveness of LAA closure was presented.  Some of the assumptions made within the study were questioned by members of the committee.  Overall, Dr. Saw indicated that LAA closure is believed to be a good alternative to DOAC or VKA.

The current clinical guidelines on LAA closure/occlusion/excision

The ESC Guidelines (2012) recommend (Class IIb) that “the interventional, percutaneous LAA closure may be considered in patients with a high stroke risk and contraindicated for long-term oral anticoagulation” (Level of Evidence B).   

The AHA Guidelines (2014) states that “the usefulness of closure of the left atrial appendage with the WATCHMAN device in patients with ischemic stroke or TIA and AF is uncertain (Class IIb; Level of Evidence B). It is also stated that “This approach is likely to have the greatest clinical utility for AF patients at high risk of stroke who are poor candidates for oral anticoagulation.

The 2014 Focus Update of the Canadian Cardiovascular Society Guidelines for the Management of Atrial Fibrillation states that “We suggest these non-approved LAA closure devices not be used, except in research protocol or in systemically documented use protocols in patients at high risk of stroke (CHADS₂ score ≥2) for who antithrombotic therapy is precluded (Conditional Recommendation, Low-Quality Evidence).

The reason why LAA closure is not 1^st line therapy is because no RCT of DOAC vs. LAA closure has been completed and that the upfront costs with LAA closure is high. The LAA closure requires physician training/skills, has procedural risks, and has procedural requirements (anaesthesiologist and TEE, and overnight admission).

Dr. Saw proceeded to address questions posed by Health Canada:

1. Given the apparent need to treat patients in Canada that are contraindicated to Warfarin, what level of clinical evidence is required to license other LAA occlusion devices in this patient population?
   1. Given that the Watchman LAA occlusion device was effectively authorized based on the totality of evidence and Meta analyses of clinical trial data and not explicitly on RCT data, could authorization of future devices be based on single arm registry data with comparisons to historic controls.
   2. Given that there are no LAA occlusion devices licensed for use in a patient population contraindicated to Warfarin, what kind of clinical endpoints or appropriate historical controls could be used?

Dr. Saw was supportive that the evidence available for the Watchman and the ACP is sufficient for licensing purposes in Canada.  Dr. Saw indicated that single arm registry data with historic controls was appropriate to assess the safety and effectiveness of the LAA Closure devices at this time.  Dr. Saw felt that the totality of the evidence was important to consider.  Dr. Saw also indicated that given the need for LAA occlusion devices in Canada, and based on the important differences in the available devices, that there would be benefit to having both devices available to clinicians.

For future devices, similar endpoints compared to those used in the Watchman Studies could be used for a patient population contraindicated to Warfarin, although in this patient population, bleed risks are less important as neither group is on OACs.  Dr. Saw felt that the manufacturers should be able to demonstrate a 50% reduction in stroke rates compared to the rates predicted by the CHA₂DS₂-VASc scale. CV death and Systemic embolization would also be factors to consider. For the population that is contraindicated to OACs, a comparison against doing nothing, where baseline stroke risk is predicted by CHA₂DS₂-VASc score, may be a sufficient historical control. Dr. Saw did mention that a RCT against DAPT may be useful but did not state it was necessary.

SAC-MDUCS committee members were asked to indicate their general agreement to this approach.  Although there was a certain level of unease, the committee did not raise any objections to this approach.

2. If post-market studies were to be proposed for future LAA occlusion devices, does the post-market study outlined for the Watchman device provide a good design for other devices, even if the patient population is somewhat different, i.e. a patient population that is contraindicated to Warfarin?

A post-market study would need to demonstrate objective evidence of stroke and systemic embolism reduction. Dr. Saw thought that pre-defined endpoints for post-market studies could be somewhat modified compared to the pre-defined Objective Performance Criteria (OPC) used in the Watchman post-market study if the study focused strictly on patients who were contraindicated to OACs. She stated that the population studied in Canada would be those contraindicated to OACs as this is current practise (and what is funded). She believes the warfarin eligible population is different and perhaps OPCs to be used for devices other than Watchman, in patients contraindicated to OACs, should be derived from large studies of that specific device in the contraindicated to OAC population.  She noted that patients contraindicated to OACs typically have significant health complexities.

Committee members strongly supported creation of registries to gather post-market data, information on device performance. It was suggested that perhaps Health Canada could work with Provinces to establish post-market registries.

3. How are the newer OACs being introduced into the treatment of patients at high risk of stroke that might also be considered for LAA occlusion?

Although it is acknowledged that novel direct oral anticoagulants (DOACs) are becoming the standard of care, there is no good clinical data to compare the use of DOACs compared to LAA closure devices. Therefore, the novel OACs are used for first-line treatment and LAA closure devices are still reserved for second-line treatment options. Committee members added that the limited use code in Ontario prevents physicians from prescribing newer OACs unless warfarin or VKAs cannot be tolerated.

A question was asked, if a manufacturer of an LAA occlusion device came forward with registry, some argument on historical control and indication different from the WATCHMAN solely in a contraindicated population, would the committee be supportive of HC issuing a license for that device with requirements for post-market follow up with different pass/fail criteria.
 
It was concluded that the biggest role is in the safety signals.  First, need to identify devices which clearly don’t meet the minimal safety criteria and after that, post market surveillance is the key. 

Dr. Parker joined the meeting.

7. Evolution of DES Technology

Dr. Marino Labinaz, Committee Member

Dr. Labinaz introduced himself and disclosed his affiliations.

Dr. Labinaz outlined the objectives of his presentation and began with an overview of Drug Eluting Stent (DES) use in Canada.  He noted that although several years ago, clinical usage was not dominated by DES, now approximately 90% of coronary stents implanted in Canada are DES, and at some centres such as the Ottawa Heart Institute, 99% of coronary stents are DES.  He highlighted that there are very few cases that remain where it is justifiable to use a Bare Metal Stent in the coronary arteries.

Dr. Labinaz provided a brief history of Drug Eluting Stents (DES). He explained that the development of DES was in response to reducing restenosis that occurred with bare metal stents. In the early 2000s, studies showed that 18 months after implantation of a sirolimus eluting stent there was persistent inhibition of neointimal hyperplasia. It was concluded that the issue of restenosis had been solved with the use of DES. Data from subsequent studies supported this conclusion. However, there was growing evidence that late stent thrombosis was occurring with DES. SCAR registry data showed that compared to bare metal stents the death rate for DES was increasing, in contrast to a plateauing of death rate of BMS patients. Meta-analysis of data from 4 randomized clinical trials (RCT) revealed that the thrombosis event rate after 1 year was higher compared to bare metal stents. Clinicians quickly learned that there are issues of bleeding with prolonged Dual Antiplatelet Therapy (DAPT) (1 year).

The severity of bleeding adverse events was also highlighted.  Although this is sometimes viewed as a minor event, data supports that there is a 3 times increase in mortality one year after a bleeding event compared to the mortality for patients who do not have a bleeding event.  Dr. Labinaz stressed that bleeding and its downstream effects are under-appreciated, and can be as significant as an MI. It is based on these considerations that work has been on-going to determine appropriate duration of DAPT for DES.  Dr. Labinaz did note, however, that there is relatively little good data to determine optimum duration of DAPT.

He indicated that investigators began to question whether they could shorten the duration of DAPT. A study was designed to investigate shortened DAPT for patients implanted with DES. Follow-up for the study was designed to be for 4 years after procedure. Results from the study demonstrated that not all stents are created equally.

The Resolute analysis looked at the effects of DAPT being interrupted within the first month. A higher stent thrombosis event rate was observed if DAPT was interrupted within the month versus a zero rate if DAPT was interrupted after 3-12 months.

Xience data showed that interruption within first 3 months had a higher event rate than interruption after 3 months. In the Optimize study, 3 months versus 12 months DAPT was investigated. The study showed no difference in Major Adverse Cardiac and Cerebral Events (MACCE) (all-cause death, MI, stroke, major bleeding) at one year. Prolonged DAPT therapy may make sense. DAPT therapy not just related to device but also related to patient.

A DAPT study was conducted to assess the benefits of 12 versus 30 months of DAPT for preventing stent thrombosis and MACE (2 primary endpoints). Almost ⅔ of stents used in this study were 2^nd generation DES. MI event rate continues to rise. DAPT beyond 1 year reduced risks of stent thrombosis and MACCE but was associated with an increased risk of bleeding/hemorrhagic stroke. Moderate or severe bleeding rates in the DAPT and placebo groups were significantly different at 2.5% vs. 1.6% respectively. In addition, the mortality rates in the DAPT vs. placebo groups were 2.0% and 1.5% respectively. Although all-cause mortality was increased in the DAPT group, a large proportion of deaths were considered bleeding related.

A FDA meta-analysis concluded there was no difference in all cause death between clopidogrel and aspirin used for 12 months or longer, versus use of both medications or aspirin alone for 6 months or less. There was no increase in cancer related events.

The Pegasus study which was not a stent related study also showed that with prolonged DAPT bleeding was significantly increased.

Based on these results the goal was to try and come up with a score to determine who would benefit the most from DAPT. A DAPT score was created. Scores < 2 saw more bleeding showing no benefit from DAPT. Scores > 2 saw significant benefit from ischemic stroke.
Guidelines for recommendation for duration of DAPT after ACS were developed in 2012. These guidelines are already out of date. An update to the original exists. Three new recommendations have been added. It also recommends individualizing duration of DAPT.

A new study has presented a challenge to the whole concept of DAPT. The study looked at monotherapy for DES patients. Prolonged monotherapy with no major bleeding was observed.

Dr. Labinaz described third generation DES:

• Biodegradable Scaffolds
• Biodegradable Polymers on Metal Stents
• Polymer Free Drug Eluting Stents

He noted that the polymer coating is an issue with DES. There are a number of shortcomings often associated with polymers during stent delivery. Durable coatings have potential for:

• continued source of inflammation
• impaired healing
• thrombosis risk

The third generation DES attempt to address issues with polymer coating.
The Absorb study showed that after 12 months vessels had become pulsatile. The ABSORB III RCT was a prospective randomized, single-blind, multi-center trial was conducted to evaluate safety and effectiveness of Absorb™ Bioresorbable Vascular Scaffold (BVS).  Individual events were low. One signal observed in the study was for patients with BVS there was a significant increase in stent thrombosis in vessel < 2.5 mm. ABSORB meta-analysis showed no difference observed compared to second generation DES.

LEADERS study was discussed. It was noted that the study’s preliminary results are out but the study is not powered well.

Dr. Labinaz listed the objectives for the Next Generation BVS Stents:

• More uniform strut distribution
• More even support of arterial wall
• Lower MCUSA (maximum circular unsupported surface area)
• Lower late stent area loss
• Higher radial strength
• Improved stent retention
• Unchanged:
  • Material
  • Strut thickness

He proceeded to address the specific questions posed by Health Canada regarding duration of DAPT following stent implantation:

a) Considerations around DAPT and shortening DAPT - modification required for the CCS Guidance statement that recommends a minimum of 1-year DAPT?

“The user is advised to take under consideration the recommendations made in the joint statement by the Canadian Cardiovascular Society and the Canadian Association of Interventional Cardiology (Can J Cardiol Vol 29, 2013) whereby all patients receiving a DES should be maintained on dual antiplatelet therapy for at least 12-months. The recommendations also emphasize the need to carefully consider the benefits and risks on an individual patient basis when choosing between Drug Eluting Stents and Bare Metal Stents; the duration of dual antiplatelet therapy and the need to counsel patients regarding the risks of prematurely discontinuing dual antiplatelet therapy. This label and all recommendations referenced are intended for information purposes only and should not be relied upon to make decisions concerning a patient's health care without further medical review.”

Dr. Labinaz listed the issues with prolonged DAPT:

• Increase risk of bleeding
• Need for oral anticoagulants
• Need for procedures and surgeries
• Interaction with other medications
• Cost

Dr. Labinaz noted that as new evidence is coming out it may alter these recommendations that were drafted in 2012.  References to contemporary guidelines for DAPT may be appropriate. Dr. Labinaz suggested that new guidelines may be based on various risk factors for bleeding and ischemic complications.  In general, 1-year of DAPT will likely be recommended.  If the patient has a high bleeding risk, 1-month duration of DAPT may be advocated as sufficient.  If the patient is at a low risk for bleeding, however, at a higher risk for embolic and ischemic events extending DAPT beyond 1-year is likely appropriate.

b) How should Health Canada address new licence applications for 3rd generation DES with biodegradable coatings and excipients as well as bioabsorbable vascular scaffolds. Should a Guidance statement be included in the labelling for DAPT?

It was recommended that the above noted guidance statement not be included in the labelling based on new data and changes that are likely to be made to the clinical practice guidelines. Dr. Labinaz presented preclinical data on the Absorb that showed pulsatility recovery and restored vasomotion at 2 years (Gogas et al. 2015). He also indicated that the Absorb III trial (Ellis et al. 2015) demonstrated non-inferiority to the Xience in an ITT analysis of target lesion failure (TLF) at 1 year. Moreover, he reported on the ABSORB meta-analysis that included 3389 patient outcomes from 4 RCTs; while no significant difference in death, MI or revascularization was reported between the Absorb and Xience, definite and probably thrombosis rates approached significance at 1.3% and 0.6% respectively. Dr. Labinaz suggested that because ST rates were similar, the potential benefit with the Absorb stent is “theoretical” normal vessel function.

c) If non-inferiority is demonstrated between a third generation DES with a second generation DES, are there clinically relevant differences (either positive or negative) that should be noted in the labelling?

It was noted that there may be differences between the devices, as highlighted with the ABSORB where there may be an increased risk of stent thrombosis, especially in smaller vessels.  It will be important to assess the data at a detailed level to determine what differences may exist.

d) Are there reasons why clinicians may not want to use some of the third generation devices in their practice, in spite of clinical data demonstrating non-inferiority?

Dr. Labinaz indicated that there may be many reasons not to use third generation stents, including:

• Inadequate Non-Inferiority Trial (e.g. too small, irrelevant endpoint, highly selected patients)
• Cost
• Deliverability
• Ease of Use
• Changes in post-procedural medical therapy
• Shelf-life, sterility, storage requirements
• Technical limitations (e.g. bifurcations, sizing)
• Availability and size catalogue

e) What should be the clinical requirements for the approval of new third generation DES; RCT vs. OPC. Do we need 2000 patients; do we need 1 and 2 year follow-up; do we need to randomize to a good currently licensed DES?

Clinical requirements for approval of 3rd Generation DES should include:

• Adequately powered randomized clinical trial with clinically meaningful endpoints against a currently approved 2nd generation drug eluting stent
• Minimum of 1 year follow-up
• Number of patients dictated by sample size calculation for an adequately powered trial (unlikely 2000 pts)
• Historical control group is inadequate since field is rapidly changing and cannot ensure that historical group is comparable

8. Baroreceptor Activation Therapy

Dr. John Parker, Guest Speaker

Dr. Parker introduced himself and disclosed his affiliations. 

Dr. Parker proceeded to discuss Baroreceptor Activation Therapy (BAT) application for the treatment of hypertension and heart failure and address specific questions posed by Health Canada.

Recently, BAT has been applied for the treatment of hypertension and heart failure. There have been several clinical trials conducted on these devices.

1. What is the mechanism of action for BAT in the treatment of heart failure?

Dr. Parker began with an overview of the sympathetic activity in Chronic Heart Failure (CHF):

• Sympathetic activation is prominent in the setting of CHF
• most available information concerns chronic CHF with reduced ejection fraction
  • Little or no information about CHF with preserved ejection fraction
• Sympathetic activation plays a critical role in CHF disease progression and prognosis

He explained that Sympathetic (neurohormonal) system activation occurs as a compensatory mechanism in patients with heart failure as a means of maintaining cardiac output, blood pressure and adequate tissue perfusion; in particular, sympathetic activation is observed in patients with heart failure with reduced ejection fraction (HFrEF). In the context of Congestive Heart Failure (CHF), increased sympathetic activation drives disease progression and worsens patient prognosis. He noted that much less is known about the role of sympathetic activation in HF with preserved EF (HFpEF; those with EF > 35%) as most available information is derived from patients with HFrEF.

Sympatholytics including beta-blockers were first introduced in the 1990s and are still considered a potent HF treatment because they block sympathetic activation. Dr. Parker described data that supported the role of sympathetic activation in heart failure. Cohn et al. (1984) showed a substantial decrease in survival probability in patients with plasma norephinephrine (PNE) levels > 400 pg/mL, as well as a significantly greater cardiac norephinephrine (NE) spillover (pmol/min) in CHF patients. Al-Hesayen, A and Parker, JD (2004) showed that both CHF and “normal LV” patients exhibited reduced LV filling pressures and increased cardiac output , infusion with the vasodilator sodium nitroprusside, did not improve either renal plasma flow (RPF) or glomerular filtration rate (GFR). Kaye et al. (1995) reported cardiac NE spillover rate was the strongest prognostic marker for survival probability, with a threshold of > 310 pmol/min.

A near tripling of sympathetic activation was reported in renal failure as well.

Dr. Parker outlined mechanisms of cardiac sympathetic activation in HF:

• Afferent
• Central
• Prejunctional.

Afferent:

Dr. Parker stated that baroreceptors on the carotid artery (“arterial baroreceptors”) monitor blood pressure continuously, with decreased pressures causing increased sympathetic outflow to the heart, kidney and periphery. Cardiopulmonary baroreceptors function in the same way. Newton and Parker (1996) reported blunted baroreflex control of cardiac sympathetic activity in heart failure patients as evidenced by a non-significant increase in cardiac norepinephrine spillover (CANESP) in heart failure patients (with elevated baseline CANESP) as compared to a significant increase in CANESP in normal patients.  Although a reduction in effect was observed over time, this reduction in effect over time has not been observed to the same extent in the Baroreceptor Activation Therapy Studies to date.

Central:

By way of example of the role that centrally mediated sympathetic activation, Dr Parker introduced the example of clonidine, a drug which reduces sympathetic overactivity centrally, and reduces peripheral vascular resistance. Azevedo et al. (1999) studied the effect of clonidine (centrally acting alpha2-agonist; decreases peripheral vascular resistance) on sympathetic activity in heart failure. They reported a substantial fall in arterial pressure as well as reduced sympathetic activity in the heart (reduced CANESP). However, because of substantial side effects, centrally acting sympatholytics are not used clinically.

Prejunctional:

Dr. Parker described prejunctional sympathetic intervention as it relates to afferent sympathetic output, and used the example of renal sympathetic denervation for the treatment of resistant hypertension. Krum et al. (2009) reported that renal denervation caused significant and sustained blood pressure reductions in hypertensive patients over 12 months, while later more definitive controlled trial results of Bhatt et al. (2014) showed no treatment effect. 

The effect of baroreceptor activation has been known for many years. In particular, stimulation of the carotid baroreceptors is known to reduce heart rate, increase vasodilation and increase natriuresis (sodium excretion) by the kidneys. As a further example Dr. Parker presented a NEJM paper from 1967 which had demonstrated reduced angina when carotid sinus nerves were stimulated.  However, translating this known effect to a safe and effective device is challenging. 

He described the CVRx Barostim neo device, and made the following comments:

• the surgical procedure is “not for the faint of heart” and not without surgical risks
• as the voltage (“dose”) goes up, there is a clear and immediate reduction in heart rate and blood pressure
• 4-year data were presented from a non-randomised feasibility study (Kroon et al, 2010), and showed a sustained reduction of SBP, but only involved 18 patients.  Data will follow-up to 5 and 6 years will be available shortly.
• Although not part of the question, he addressed findings from the Rheos hypertension pivotal trial, which was designed as a double blind crossover trial, and which missed its primary endpoints: difference in 6 month reduction in SBP was < 20% goal, sustained effectiveness was not significantly different between groups at 12 months, and short term (30 day) procedure adverse events were higher than expected and missed OPC of 82% (4.4% permanent nerve injury; 4.8% transient nerve injury; 4.4% general surgical complications; and 2.6% respiratory complications.)
• For heart failure specifically, Dr. Parker presented a positive proof of concept study in dogs. Sabbah et al. (2011) reported small but significant reductions in LVEDP, increase in cardiac output, increase in LVEF%, and decreased PNE in 14 dogs compared to 6 control dogs.

2. What clinical evidence for the safety and effectiveness would be considered to be sufficient for premarket licence approval of the intended use in patients with heart failure?

Dr. Parker described the BAT for HFrEF study (also known as the XR-1 Randomized HF Study) - a randomised BAT versus medical therapy trial with 146 patients randomized 1:1.  He stated that although the study was primarily exploratory, there is a primary effectiveness objective specified, namely to determine whether the Barostim neo produces an increase in LVEF at 6 months post implant.  The study had no blinding and there was no sham or crossover included in the study.  The patient population is fairly sick, with Class III HF, but have relatively low BPs, i.e. they are “well treated”.  Dr. Parker described the inclusion/exclusion criteria as “straightforward”.  The implants were well tolerated, and it appears the clinicians were well trained in technique.  Dr. Parker found the reported complication rates acceptable, and indicated BAT therapy did not produce hypotension in patients. BAT was found to reduce NYHA HF class. QoL increased, and N-terminal pro Brain natriuretic peptide (NT-proBNP) went down.    
Key point: Dr. Parker noted that NT-proBNP is the only truly objective endpoint in this trial. He also acknowledged that most regulatory agencies focus on “hard endpoints” including mortality and HF related hospitalization and mortality.   This is notwithstanding the fact that the reduction was very small, and the treatment group had a higher baseline; however, he still felt that it indicated that there is some benefit to the treatment.
 
Dr. Parker noted that the effect of BAT on LV Ejection was not statistically significant. He acknowledged that the 6-minute walk test is not overly meaningful.

Dr. Parker cautioned that in the MOXCON drug trial, that acted on a similar mechanism; there was an observed increase in cardiac mortality in a large trial. A total of 71 patients were implanted with the BAT neo.  The neo received CE mark based on this trial.

3. What are the most appropriate clinical endpoints when studying the effectiveness of treating heart failure?

Dr. Parker recommended that a combined endpoint of heart failure hospitalisation and mortality “has to be the goalpost”, and that this is typically what is used for heart failure drugs.  Other endpoints like exercise tolerance are not appropriate.  A larger trial with the recommended endpoints is required.  A sham trial is not required if hard endpoints are used (e.g. death, or possibly, blood markers), although a crossover trial design might be beneficial.

4. What is the appropriate target population for this type of device?

Dr. Parker suggested that appropriate target populations include patients with both reduced and preserved ejection fraction, because the latter group are often hypertensive and may stand to benefit from this therapy.

Members commented on the small BAT-induced NT-proBNP reductions. Question was asked whether Dr. Parker believed the reduction of this biomarker was a reasonable surrogate endpoint. He thought it was, and added that he believed the 6MHW distance was not a good test.

Dr. Parker was asked about the effects of afferent neural traffic from the contralateral side; he was unsure but referred to carotid massage given only on the right side. There was a question about the potential for BAT-induced bradycardia and Dr. Parker indicated he was unaware of any such reports.

Dr. Parker reiterated that a sham study design is not required if hard endpoints (death, or possibly blood markers) are used. He was also asked about long-term remodelling and functional improvement as a result of BAT and reiterated that a small +5% increase in ejection fraction was reported without any other significant changes.

Dr. Parker was also asked if there were any metabolic studies done in relation to BAT; he was not aware of any. Dr. Parker noted that in endarterectomy, nerve damage may be observed at a rate of approximately10%.  The rate seen in this study was only 5%, although the amount of dissection would be less for implanting the BAT neo compared to endarterectomy.

Finally, Dr. Parker suggested that, based on the experience with beta blockers, we might expect to see a 5% (absolute) change in ejection fraction.  However, this has yet to be shown.

9. IVC Filters

Dr. William Geerts, Committee Member

Dr. Geerts introduced himself and disclosed his affiliations. 

Dr. Geerts was asked to discuss use of Inferior vena cava (IVC) filters and address specific questions posed by Health Canada.

Inferior vena cava (IVC) filters are medical devices designed to be inserted into the IVC to trap emboli and prevent pulmonary embolism (PE). There have been serious concerns about the safety and effectiveness of these devices and new RCT data has shown little to no benefit using the technology in certain patient populations. Health Canada has been working on these issues for years and has prepared the following recommendations:

1. Follow the appropriate clinical guidelines on the use of IVC filters. Currently, Health Canada considers that the indications for use of IVC filters recommended in the 9th ACCP guidelines and AHA/ACC are appropriate.
   1. In patients with acute proximal DVT of the leg and contraindication to anticoagulation, we recommend the use of an IVC filter
   2. In patients with acute PE and contraindication to anticoagulation, we recommend the use of an IVC filter

Note that the FDA provides the following indications:

For the prevention of recurrent pulmonary embolism via placement in the vena cava in the following situations:

1. Pulmonary thromboembolism when anticoagulants are contraindicated
2. Failure of anticoagulant therapy in thromboembolic diseases
3. Emergency treatment following massive pulmonary embolism where anticipated benefits of conventional therapy are reduced
4. Chronic, recurrent pulmonary embolism where anticoagulant therapy has failed or is contraindicated

b. Make and implement appropriate follow-up plan in cooperation with patients to ensure compliance of patients to the follow-up plan.

c. Take necessary measures to improve the retrieval rates of implanted retrievable IVC filters when not needed.

d. It is strongly recommended that the patients who receive the procedure should be automatically scheduled for retrieval consultation at time of placement of an IVC filter. This information should be presented on the labeling for the IVC filters.

Health Canada would like to seek feedback from the SAC members regarding these recommendations.

Dr. Geerts began with the recent news of the two class action lawsuits in Canada regarding IVC filters. He believed that these were the first two in Canada, although many lawsuits have been filed in the US.

He indicated that he does not believe there is widespread IVCF overuse in Canada, but indicated that he is aware of Canadian cases in which the use of an IVC filter was not necessary or appropriate. He expressed the view that the most worrisome aspects of IVC use are filter fracture and embolization and failure to remove filters when they are no longer necessary.

Dr. Geerts provided an outline of his presentation:

• Types of filters
• Epidemiology of IVCF use
• Indications
• Evidence from clinical trials
• Complications
• Guidelines & national policies
• Summary and suggestions

He proceeded to discuss types of IVC Filters:

• Permanent
• Temporary.

Temporary filter may be tethered or retrievable. He indicated that retrievable designs are the most common, with the (Cook Medical) Celect and Gunther Tulip models being the most common in Canada.  He commented that there may be a small clinical role for the “tethered” type designs, but none are licenced in Canada at this time.

Regarding indications, there are 2 categories:

• “Therapeutic” for patients with venous thromboembolism (VTE) in which anticoagulation is contraindicated, or in addition to anticoagulation) and
• “Prophylactic” for patients without VTE, and presumed to be at high risk for pulmonary embolism from trauma, bariatric surgery, or orthopedic surgery.

Based on US data (Stein, 2004; Smouse, 2010; Stein, 2011) IVC filter placement has increased exponentially over the past 40 years, with the largest per cent increases attributed to the “prophylactic” indication group. Therapeutic IVC filter use was reported to increase 42% between 1998 and 2005, while prophylactic use increased 157% during this same period.

Dr. Geerts also discussed the large variation in IVC filter use for VTE by hospitals (White, 2013; US data) and by different countries where  the rate of IVC filter placement in US patients was 25x greater than in German, French, British, Italian and Spanish patients combined (Lee, 2015).  He suggested this kind of variation cannot be explained based on clinical factors, and presented possible reasons for the US increase: expanded indications, ease of placement, more experienced inserters, removable filters, “defensive” medicine, and possible financial incentives. He emphatically stated that the exponential increase in use is not the result of additional evidence for safety or benefit. While he was not aware of any Canadian use data, he did provide information on use of IVCFs at Sunnybrook (his institution); where they inserted an average of only 21 IVCFs per year between 2006 and 2015 for all indications (none were for a prophylactic indication). Furthermore, in trauma patients, Sunnybrook placed an average of 4 IVCFs per year between 2000-2015 (none prophylactic) and 76% were removed during the trauma admission.

Additional data from the Ottawa Hospital (Duffet, 2014) was included.

Dr. Geerts summarised his interpretation of the only 2 RCTs available which purport to support the use of IVC filters.

Decousus (1998) described the PREPIC study which included 400 patients with DVT and treated for increased PE risk, and randomized to either anticoagulation (AC), or AC+IVC filter. Pulmonary embolism (PE) at Day 12 was significantly lower in the IVC filter group. Dr. Geerts asserted that this study “has been grossly misinterpreted” because they included asymptomatic PEs and mostly unadjudicated deaths (although attributed to PE) in the endpoint, thus overstating the effect of the use of filters. Similarly, the PREPIC-2 trial (Mismetti, 2015) enrolled 399 patients in 18 French centers that were randomized into AC, or AC and retrievable IVC filter groups. No significant differences in PE at 3 months, or fatal PEs were reported. Dr. Geerts indicated there was a numerical increase in PE rate in patients implanted with IVCFs (6 vs 3), and Fatal Pes (6 vs 2). He also highlighted that access site hematoma, filter thrombosis, unsuccessful retrieval and non-retrieval of filter rates were 2.6%, 1.6%, 6.7% and 21% respectively. Finally, he noted that systematic reviews have found no Level 1 evidence to support the use of filters for prophylactic indication (i.e. without a VTE).

Dr. Geerts presented information on the FDA requirement for additional IVCF clinical data from manufacturers, who were given a choice to either participate in the PRESERVE study, or to conduct their own post-marketing surveillance (“522” studies). Starting in Nov 2015, the first patient entered into the PRESERVE study, which is a multi-device trial intended to examine current practice and to evaluate safety and effectiveness of these devices.  A total of 300 patients per device type were required, with up to 24 months of follow up. Manufacturers were also given the option to combine data from multiple devices to enhance data collection.

Dr. Geerts then provided a detailed analysis of IVC-related complications. Key points made were that PE (and occasional fatal PE) still occurs with a filter in place (i.e. filters trap larger clots, but not all clots), and that long-term fracture and embolization are big problems with these designs.

He estimated that IVC filter use is associated with a minimum major complication rate of 3-5% and commented that this was greater than the complication rate from treated PE or from major PE in high risk trauma patients. He showed the results of a study that showed the risk of IVCF fracture in patients, followed over time with imaging, was 40% at 5.5 years. He shared that fractured filter arms could migrate and perforate the right ventricle, leading to tamponade, and can penetrate other organs especially the aorta, duodenum and retroperitoneal space.

Other adverse consequences he attributed to IVCF placement included: unnecessary delays in initiating AC prophylaxis or therapy in patients who have a filter and potential risk of complications during pregnancy and exercise.

He noted that it is very likely that complication rates attributed to PE/TE are less than the complication rates due to filter use, even though the filters are intended to reduce PE.

Dr. Geerts presented information from the American College of Chest Physicians, which recommend against thromboprophylaxis of IVC filters in trauma patients or those with spinal cord injuries (Gould, 2012), and against the use of IVC filters in patients with acute DVT or PE who are being treated with anticoagulants (Kearon, 2016). He also highlighted the FDA’s concern over the retrieval of IVC filters after the risk of PE/TE has passed. Their 2014 recommendation is, “The FDA recommends that implanting physicians and clinicians responsible for the ongoing care of patients with retrievable IVC filters consider removing them as soon as protection from pulmonary embolism is no longer needed”.

The recommendations from Thrombosis Canada were presented. Dr. Geerts closed with a recommendation for a single indication for IVC filter use (recent proximal DVT plus an absolute contraindication to full anticoagulation), and also provided a summary of the guidelines in place at Sunnybrook.

With respect to the Health Canada proposal, he provided the following comments:

• He thinks that while the AHA/ACC guidelines provide some good advice, they are now 5 years old and expand the indications inappropriately, therefore he suggests deleting reference to them (identified in Section  a of preamble above).
• He agreed with Indication 1 (identified in Section  a - i of preamble above)
• Indication 2, possible, but it is not what he or his institution does (identified in Section a - ii of preamble above). He suggested the following wording: In patients with acute PE and contraindication to anticoagulation, use of an IVC filter can be considered.
• He does not think that patients should take the primary responsibility for appropriate follow-up - the institution that places the filter should assume this responsibility 
• He emphatically agrees that necessary measures should be taken to improve retrieval rates of implanted filters
• He feels that the retrieval should be scheduled at the time of placement.
• Regarding the FDA indications (identified in Section  a of preamble above) he noted that Health Canada may want to comment that there is no direct evidence to support any of these indications and that Health Canada does not endorse this list of indications.

Dr. Geerts provided some final thoughts:

• It would be important to have Canadian data on IVCF use, indications, types of filters, retrieval rates, complications, and costs in the Canadian environment.
• Retrieval is much easier at 6 days than at 6 months.  His institution performs retrievals before patient is discharged.
• Hospitals should have written policies on IVC filter use
• New filter approvals must present lots of safety data, and must be subject to post approval data provision requirements.

Discussion took place about how best to share this information, in particular with interventional radiologists who are implanting these devices. It was noted that Health Canada has difficulty with information communication as the majority of manufacturers have been unwilling to make label changes. Different means of communication were discussed (i.e. patient implant cards) or a possibility to reaching out to radiologists and suggesting their involvement in a solution. Dr. Geerts thinks while it’s too little too late to introduce patient implant cards, he feels that engaging radiologists would be the main next step. We could certainly identify leaders in radiology who would be willing to participate.

Dr. Ducas noted the differences in IVCF implantation rate between Sunnybrook and Ottawa Hospitals, and suggested if they could get together to examine use differences and compare post-market issues.  He also suggested if Health Canada would network with provincial representatives to work on a registry.

Dr. Geerts suggested that a position statement from HC may not only highlight the importance of device removal, but also drive the communication process. Dr. Geerts mentioned that he is involved with the CIHR-funded CANVECTOR group, which is involved in research, knowledge translation, patient involvement and education; he suggested this as a potential platform for collaboration.

Dr. Geerts was generally not supportive of IVCF use, with the exception of a single indication: recent proximal DVT plus an absolute contraindication to full anticoagulation. He was in agreement with indication a i), and would not refute a ii), but this was not practiced at his hospital.

Furthermore, he felt that patients should not have to take the primary responsibility for appropriate follow-up, retrieval rates are in need of improvement and retrieval should be scheduled at time of placement. He also urged that hospitals have written policies in place.
 
With respect to regulation, sufficient safety data should accompany IVC filter applications and post approval data should be required.

A HC position paper and/or statement were suggested as a potential vehicle for communicating the importance of patient follow up and device removal.

10. Next Steps, Closing Remarks and Adjournment of Meeting

Dr. John Ducas, Committee Chair

The Chair thanked committee members and Health Canada staff for their participation and valuable input. 

Members will be canvassed to select a date for the next meeting, likely in the late fall of 2016.

Meeting adjourned.