Scientific Advisory Committee on Medical Devices used in Cardiovascular Systems - Record of Proceedings – November 25, 2016

Committee Members: John Ducas (Chair), Renzo Cecere, Eric Cohen, John Ducas (Chair), Christopher Feindel, Marino Labinaz, Brent Mitchell, Joaquim Miró, Barry Rubin, John Webb, Raymond Yee, John Webb (via teleconference/WebEx)
Regrets: Alan Menkis
Invited Guests: Mark Peterson, Kong Teng Tan
Health Canada Representatives:
Office of Science: Hripsime Shahbazian, Larissa Lefebvre,
Medical Devices Bureau (MDB, TPD): Kimby Barton, Kevin Day, Ian Aldous, Ben Elliot, Ian Glasgow, Jianming Hao, Karen Kennedy, Mark Korchinski, James McGarrity, Bisi Lawuyi, Roy Masters, Catherine Milley, Chris Schmidt, Maurice Sylvain, Lanyi Xu
Bureau of Cardiology, Allergy and Neurological Sciences (BCANS, TPD): Mick Gelsena, Xiaobing Guo, Tiamo Li
Regulatory Operations and Regions Branch (RORB): Collin Pinto
Marketed Pharmaceuticals and Medical Devices Bureau (MHPD): Melanie Bousquet, Patrick Fandja, Chengwen Ren

Abbreviations used in this record:

AI:

aortic insufficiency

AV:

atrioventricular

BiV:

bi-ventricular

BMS:

bare metal stent

CABG:

coronary artery bypasses grafting

CRT:

cardiac resynchronization therapy

DCB:

drug coated balloons

DEB:

drug eluting balloons

DES:

drug eluting stent

EAO:

effective orifice area

ICD:

implantable cardioverter defibrillator

ISR:

in-stent restenosis

LVAD:

left ventricular assist device

LVEF:

left ventricular ejection fraction

LV:

left ventricle

MACCE:

Major Cardiovascular and Cerebrovascular Event

MDB:

medical devices bureau

NYHA:

New York Heart Association

pAVD:

paced Atrial-Ventricular Delay

pAVI:

paced Atrial-Ventricular Interval

PCI:

percutaneous coronary intervention

POBA:

plain old balloon angioplasty

RCT:

randomized clinical trial

SAC-MDUCS:

scientific advisory committee on medical devices used in the cardiovascular system

SAP:

special access program

sAVD:

sensed Atrial-Ventricular Delay

sAVI:

sensed Atrial-Ventricular Interval

SAVR:

surgical aortic valve replacement

SFA:

superficial femoral artery

STEMI:

ST-elevation myocardial infarction

STS:

The Society of Thoracic Surgeons

TAVI:

transcatheter aortic valve implantation

TAVR:

transcatheter aortic valve replacement

TMVR:

transcatheter mitral valve replacement

TLR:

target lesion revascularization

TPD:

Therapeutic Products Directorate

TVR:

target vessel revascularization

1. Opening Remarks & Welcome

Dr. John Ducas, Committee Chair

Dr. Ducas opened the meeting, welcomed the committee members and guest speakers. Kimby Barton, the current Interim Director of the MDB and Senior Executive Director of the Director General's Office 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. She summarized issues that were brought to the committee for discussion and recommendations. She noted 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 acknowledged the committee on its continued support and noted that their past recommendations have been helpful in numerous license evaluations. She concluded by wishing the committee members and participants a productive meeting.

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.

He invited Hripsime Shabazian to address the next topic.

3. Review of Terms of Reference, Membership and members Affiliations and Interests

Hripsime Shahbazian (Office of Science, TPD)

Hripsime Shahbazian informed the committee members regarding renewal of membership, as per the SAC-MDUCS Terms of Reference (ToR), which requires that memberships should be reviewed periodically. She noted that the Office of Science conducted membership renewal in 2014, and after 2 years it is time once more to revise and update membership for the committee. At this time HC would like to reaffirm member's continued interest in participating as a member or an ad hoc member of the SAC-MDUCS.

The Affiliations and Interests Declaration Forms need to be revisited as well. This form requests information from member's current situation as well as history regarding their affiliations and interests over the last five years. Once completed, the information provided in this declaration will be reviewed and then used by Health Canada to produce a "Summary of Expertise, Experience, and Affiliations and Interests." Every member will have an opportunity to review, amend and/or approve their own summary which will be posted to Health Canada's website as required under section 3.0 of the Health Canada Policy on External Advisory Bodies (2011) to promote our Branch's commitment to be open and transparent regarding the membership of its advisory bodies.

In addition, members will be asked to provide an updated Biography. Each member will be asked to review their biography as it appears now on Health Canada web and indicate whether it is acceptable as written or whether changes are required to this version.

Furthermore, Mrs. Shahbazian noted that before each committee meeting the chair will ask the members/invited experts to provide their consent to the recording of their views and opinions in the Record of Proceedings (RoP) and to the publication of the RoP on Health Canada's Web site. Once a draft RoP is prepared by the Secretariat it will be circulated to members for review, and final approval by the Chair and Executive Secretary. All members/invited experts that attended the meeting will be given the opportunity at the time of reviewing the draft RoP to ensure that their views are properly recorded. The final RoP will effectively summarize the proceedings to reflect the advice offered. The Secretariat is responsible for the distribution of the RoP. This process will be reflected in the revised ToR and will be circulated to the members for their review and acceptance. All members supported this requirement.

Mrs. Shahbazian thanked members for their continued support.

4. 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:

• Bioabsorbable Vascular Scaffolds: good overview of both the strengths and weaknesses of the technology were noted. Health Canada will continue to monitor the device performance as the data becomes available.
• Chronic Heart Failure and Baroreceptor Activation Therapy: background information was valuable in determining if the technologies should be approved for the Canadian market.
• Inferior Vena Cava Filters: Health Canada has received positive feedback regarding the Notice including the focus on increased patient follow-up.
• Leadless Pacemakers: was informed of clinical trials and adverse event assessments with other leadless pacemakers, and determined appropriate indications for use and appropriate clinical endpoints.
• LAA Closure Devices: Health Canada continues to monitor and access the data as it becomes available.

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

5. Various techniques to improve cardiac resynchronization therapy (CRT) responder rates

Dr. Raymond Yee, Committee Member

Dr. Yee disclosed his affiliations and proceeded to address the questions posed by Health Canada.

OBJECTIVE: Increasingly, manufacturers of CRT systems are trying to introduce new features that appear to target improving the CRT responder rates. Frequently, the clinical data is relatively weak in terms of demonstrating improved clinical outcomes that are associated with clinically relevant endpoints with these new features. Health Canada requests input and recommendations on how these new features should be reviewed and what level of evidence should be required prior to issuing a licence to these types of features for CRT systems.

1. The idea makes sense and most device manufacturers now include some functionality to try and optimize AV delay in CRT systems. However, there exists meta-analysis and other data that don't support the clinical benefit of AV delay optimization. Do the experts believe this represents an effective treatment strategy?

   Experts agree that sAVi (sensed AV interval) & pAVi (paced AV interval) programming are important for optimal cardiac function in HF patients, and VVi (Ventricle to Ventricle interval) programming may be valuable in partially overcoming problems of suboptimal LV lead placement. However, it should be remembered that the important AV intervals pertain to the left heart so that the timing of LA and LV contraction are most important. CRT non-responder rates continue to be around 35% with ranges reported in the range of 15% to 45%. Experts desire but are skeptical of finding an algorithm that improves CRT response rates, and the magnitude of benefit may be limited since CRT response is determined by multiple factors. Optimization algorithms are important in setting programming boundaries, and patients require continual follow-up as their hearts remodel. Optimization algorithms should be efficacious and safe; desire ones that are more simple, less costly to implement and less resource intensive than ECHO optimization but that may not be achievable. Continuous or auto-adjusting algorithms are preferred; continuous adjustment will be required as the patient remodels. Hemodynamic-based algorithms with closed loop feedback preferred.

2. What methods and features of existing CRT systems are being used to try and improve CRT responder rates? Is there data to support these improvements?

   Three strategies for increasing CRT response rate were presented: patient selection, procedural modifications, and post-implant care. Optimization algorithms are involved in the post-implant phase only.

   Patient characteristics associated with greater responder rates, based on RCT studies include:

   • women have a higher response rate than men (men still respond) due in part to greater proportion of patients with non-ischemic cardiomyopathy
   • extent of focal myocardial scar
   • prior history of HF hospitalization
   • left bundle branch block (patients with right bundle or non-specific interventricular induction delay showed no benefit in the subgroup analysis)
   • greater QRS duration/width (i.e. > 150 ms – greater degree of myocardial dysynchrony) (but the RAFT study showed that patients with paced QRS complexes at ≥ 200 msec tended to not respond)
   • left atria size (larger LA did worse)
   • atrial fibrillation (patients who stayed in sinus rhythm did better)

   Smaller trials have identified predictors of poor CRT response including:

   • poor RV function
   • poor renal function
   • greater myocardial scar burden

   The presence of arrhythmias has an effect on ability to pace. The devices are programmed to pace in a range of between 50-60 beats/min to as high as 120-130 beats/min depending on the patient. The age and functional capacity of a patient, at baseline, will determine the programmed pacing. Generally, the less CRT pacing being delivered (pacing burden) by the device, the less likely the patient will respond (analogous to drug dosage in pharmacologic therapy).

   Atrial fibrillation is a major clinical problem limiting response, with the two major consequences being: loss of atrial systole, and higher heart rate (which decreases the ability to pace).

   Two treatment strategies are available for AF patients (insufficient scientific evidence regarding which is the best):

   1. Maintain sinus rhythm (rhythm control) with antiarrhythmic drugs or AF ablation, and
   2. Ventricular rate controls by AV node blocking drugs or AV node ablation.

   The clinical benefit of CRT in patients with permanent AF is currently undergoing assessment by randomized clinical trial: RAFT Perm-AF.

3. What have we learned and what data is there to support the possible identification of responders or non-responders prior to implantation.

Definition of CRT responder

It is important to note that there are various definitions of CRT response and they do not necessarily correlate:

• Clinical or subjective using criteria such as decreased HF symptoms, decreased QoL Score, and change in NYHA functional Class > 1
• LV remodeling or objective using criteria such as; absolute change in LVEF >= 5% or change in LVESV >= 10-15%
• clinically relevant endpoints including total or cardiovascular mortality, HF hospitalization; and, combined clinical composite score.

Response/non-response to CRT is usually determined 3-6 months post-implant. Suboptimal AV timing has been identified in majority in many of non-responders, but there are other contributing factors including arrhythmias.

The location of the pacing leads is another important factor determining likelihood of CRT response and is potentially within the control of the physician. Dr. Yee suggested that pacing the LV lead near the apex has been shown to have poorer outcomes (MADIT-CRT). However, there may be individual patients in whom this may be the best or only placement option. Pacing basally in posterolateral or anterolateral wall segments generally is better but lead placement is limited by coronary venous anatomy. Therefore, procedure-based strategy for trying to increase response rate involves optimal lead placement, particularly the LV lead. Approaches include targeting LV leads to avoid myocardial scar and to pacing at the site of latest LV wall segment activation (Q-LV).

Myocardial scar slows conduction of electrical wavefronts to the rest of the LV so pacing within scar would not be expected to synchronize wall segments as effectively. There is much less information about the importance of RV pacing site and RV scar. The MAPIT-HF trial, which was a smaller trial, found that site of RV pacing is also a factor in determining response. Patients whose pacing leads avoided scar in either ventricle showed the greatest likelihood of responding (80% response rate) versus 54.5% responded the RV lead landed at the site of RV scar and 25% if the LV lead was located at the site of scar. The response rate was zero if both leads happened to land in scarred myocardium.

Results from two randomized clinical trials (TARGET & ImagingCRT) indicate that imaging-directed lead placement increases CRT response. TARGET was the first trial to test the hypothesis. In half the patients, LV lead position was guided by echocardiographic radial strain tracking. Study details include:

• 220 patients, randomized, primary endpoint was echo defined response of greater or equal to a 15% decrease in LV end systolic volume.
• Control: LV lead was placed by the surgeon's preference
• Targeted: LV lead was echo guided and placed away from scaring
• Primary endpoint response: was LVESV decrease by ≥ 15% from baseline
• Response rate in the control group was 55% and was 70% in the targeted lead placement group
• Cost is minimal in terms of procedure time (no significant increase is procedure time)

In addition, it's best to place the lead at the latest activated site defined either mechanically or electrically. Most people do it electrically by measuring the QLV interval (during intrinsic supraventricular rhythm) which is defined as the time from the initial deflection of the QRS on the surface ECG to local intrinsic activation at the LV stimulation site. Furthermore, the greater the degree of narrowing of the biventricular paced QRS (to the narrowest width), the greater the likelihood of response.

Alternative pacing strategies are also being tested. For example, biventricular pacing has been compared to His bundle pacing in a crossover design in two small studies. Patients were followed for a year, and His pacing had similar effects on EF, NYHA Class, QoL, or 6MHW.

Discussion regarding area of scar:

There was a comment, "Dead meat don't beat", made by the committee, which raised the question, "Is scar burden responsible for the lack of response?" Dr. Yee responded that it is an important factor but it is but one of many. He indicated that the more scar [present], the less likely you'll see a response, and that this was supported by MRI data in CRT patients.

4. What are the risks and benefits of multi-site pacing compared to standard Bi-Ventricular (BiV) pacing. The discussion might include:
1. a quantitative or qualitative comparison of the safety and effectiveness of Multi-site versus BiV pacing;

   CRT-P (CRT-Pacemaker) is a BiV pacemaker. CRT-D is the same as a CRT-P with the addition of implantable cardioverter-defibrillator (ICD) function. 95% of the CRT-D pacemakers are implanted. However, the reason a CRT-D is not implemented has to do more with patient preference – they do not want to receive a shock and die.

   Dr. Yee indicated that there are many different versions of multisite pacing and they need to be distinguished. Some versions may be more efficacious than others and it is important not to lump them all together. He suggested that multisite pacing may potentially increase CRT responders, because it would allow physicians to choose the best combinations of pacing sites. He identified the risks as:

   • increased procedure complexity and time, radiation exposure
   • lead dislodgments, and lead-lead interactions
   • potential for increased battery energy consumption

   The benefits include: improved CRT response reduced HFH (Heart Failure Hospitalization) and mortality. He indicated that at this time there is insufficient evidence to support this strategy.

   Four trials have compared these approaches (TRIP-HF, Lanerczyk, Rogers, Pappone). All have small number of patients and are unicenter or multicenter, have serious design limitation (e.g. crossover studies lack a wash out phase), different endpoints are used in each, and too many different alternatives are evaluated (2 RVs leads, 2 RV + 2 LV leads) to interpret. Also, patients with prior pacemakers were included.

2. evidence that Multi-site pacing could provide a quantifiable clinical benefit compared with BiV including the support from available data from pivotal clinical trials;

   It is important to distinguish between multipoint pacing (MPP) using electrodes on a single LV lead in one coronary vein versus electrodes on separate LV leads in two different coronary veins.

   The Multi-Point Pacing Technology (MPP) trial (only in Abstract Form – not yet published) is an industry sponsored, multi-centered, randomized study, comparing multipoint pacing 2 sites in the LV combined with RV versus standard BiV pacing using a single quadra polar lead, longitudinally along the LV. The primary safety endpoint was freedom from system-related complications at 9 months. The primary efficacy endpoint was non inferiority of response of MPP at 9 months compared to 3 months of BiV pacing. Safety response (93.2% freedom from system-related complications) met the safety endpoints. Effectiveness endpoint showed no substantive difference between MPP and standard BiV pacing, although Dr. Yee pointed out that the study showed a trend towards multipoint pacing having a worse outcome but non-significantly; the available data did not point to a benefit, but only ensured there was no harm.

3. the clinical implications of reducing battery life with Multi-site pacing

   Reduced battery longevity results in more generator replacements during the device lifetime and more exposure to infection, which translates to higher lead extraction risk.

   The potential impact of MPP on battery life was discussed. It was suggested that turning the MPP feature on results in a loss of one year of battery life. There was a suggestion that manufacturers should develop and provide software that includes a warning function to notify the user of the battery life that would be lost from activating multipoint pacing.

   Other alternative CRT pacing strategies were discussed:

   • His bundle pacing (at the base of the His bundle) may correct LBBB.
   • 2 separate LV leads vs. one IS-4 lead, seems to be no worse.
   • LV Endocardial pacing is more physiologic and provides homogeneous LV activation. The ALSYNC trial included a single cohort to assess the safety and performance of atrial trans-septal tools and LV endocardial BiV pacing. It included 138 non-responder patients for which CRT was indicated, as well as those who failed or were unsuitable for conventional LV lead placement. A 89.4% success rate was reported, and at 6 months NYHA improved in 59% of patients. Also, freedom from complications at 6 months f/u 82.2%
   • Q-LV interval mapping, and CRT response was evaluated as part of the SMART AV sub-study, that included 426 of the 846 patients in the main study. Q-LV intervals greater than 95 ms were associated with significantly lower LVESV and LVEDV, and significantly higher LVEF, but lower QoL scores.

   Optimizing Atrial-Ventricular timing was presented as another alternative CRT pacing strategy, as LA and LV systolic timing is critical to systolic and diastolic function.

5. Discuss when these features might be employed for patients; are these features mostly likely to be used only once someone is identified as a non-responder?

   This depends on the nature of the technology (hardware vs software solution) aimed at improving CRT response and the balance of benefit vs. risk. If hardware based, it would be best implemented in every patient at the time of implant since not easy to predict who will be a non-responder, and it might also make responders even better if applied to everyone. However, it may need to be reserved for non-responders only if risk and costs of applying the technology are high. If software based, it may be reasonable to allow these features to be turned on in cases where the patient has been identified as a non-responder.

   How and when are you confident that someone will not respond to standard CRT?

   CRT response is usually determined at 3 or 6 months post-implant. If the patient is not responding to CRT therapy by 6-months, it [response] is unusual, but not impossible to be a "late responder".

   If features are only to be used with non-responders, does this change the evidence required to demonstrate improved responder rates?

   No. The definition of response is problematic. Having said that, however non-response is defined, the need and type of evidence that a new technology converts non-responders to responders remains the same.

   Dr. Yee indicated that aftercare of patients with CRT is important by trying to optimize response in patients.

   • Watch for atrial fibrillation and try to control atrial fibrillation to maximize pacing algorithms to try and mimic optimization.
   • Timing cycles based on the right side (pacing is typically right heart based) whereas controlling LA/LV systolic timing is critical to systolic and diastolic function.
   • Use echo to optimize: echo optimization involves trying to optimize the AV delay so that the E wave (represents the ventricular filling) and the A wave (represents the atrial systole) are optimally separated instead of overlapping. This is done to ensure optimal filling of the ventricle between contractions.

   Most centers do not perform routine echo optimization on all CRT patients since it is time consuming and expensive (ECG based timing is more common practice). When in doubt the AV delay is set to a short number.

   Optimal LA-LV timing depends upon the following:

   • RA to LA conduction time
   • AV nodal and RV to LV conduction time
   • Body position / preload
   • Heart rate / activity level
   • Intra-individual and inter-individual factions

   Optimal RV-LV timing depends on the following:

   • Severity of antegrade LV conduction delay (LBBB and intra-myocardial)
   • Location and severity of myocardial scar
   • Epicardial vs. endocardial pacing site (anisotropic conduction)
   • Conduction time between RV and LV pacing electrodes

   Dr. Yee summarized the current practices and thinking.

   1. If the substrate or lead placement is poor, AV and VV adjustments will only marginally improve response rates.
   2. Hemodynamic based adjustments are preferred to auto-adjusting algorithms.
   3. Routine echo based AV, VV timing optimization is not recommended or practiced, but is reserved for non-responders.
   4. sAVD (sensed AV delay) and pAVD (paced AV delay) are set to relatively short values (E.g. 100 and 130 msec respectively)
   5. BiV pacing % is monitored and attempt are made to maximize the amount of BiV pacing.
   6. aCRT (adaptive) has advantages of automation, and reduced battery energy consumption.
   7. Some work to better measure CRT response is being done with accelerometers embedded within leads and with impedance measures in the leads.

Dr. Peterson, Dr. Tan and Dr. Webb joined the meeting.

6. Transcatheter Aortic Valve Replacement (TAVR)

Dr. Marino Labinaz, Committee Member

Dr. Labinaz disclosed his affiliations and proceeded to address the questions posed by Health Canada.

OBJECTIVE: With more clinical data becoming available for the outcomes associated with TAVR systems, the next major regulatory assessment will involve considering the use of TAVR in intermediate risk patients. To date, there is not a lot of long-term data associated with TAVR systems and Health Canada is concerned that issues such as valve durability and possibly valve thrombosis could present problems as the technology is introduced into healthier patient populations. It is also less clear as these intermediate risk patients could be reasonable surgical candidates, what role the Heart Team will play in the decision making process. Health Canada is looking for recommendations and clinical issues to consider as we anticipate having to assess licence applications for this intermediate risk patient population.

To date, only two families of TAVR systems are licensed within Canada (Sapien and CoreValve). Licensing has been limited to high risk patients for these devices. In some cases, approval for femoral access routes have allowed for a broader patient population compared to alternate access routes such as direct aortic or trans-apical.

Clinical trials are currently underway to generate clinical information required to support expanding the indications to include a healthier target population who are currently typically treated with surgical bioprosthetic valves. Recommendations and comments on the following issues are requested:

Dr. Labinaz presented the CoreValve US Pivotal Trial and the PARTNER 2A Trial as some background on use in intermediate risk patients.

The CoreValve Trial

The CoreValve Trial was designed as a high risk trial but unintentionally became an intermediate risk trial due to the patient characteristics (i.e. STS scores of 7.4 ± 3.2) of the subjects enrolled. Inclusion criteria were: New York Heart Association (NYHA) functional class II or greater; severe aortic stenosis (AVA ≤ 0.8 cm2 or AVAI ≤ 0.5 cm2/m2 AND mean gradient > 40 mm Hg or peak velocity > 4 m/sec at rest or with dobutamine stress echcocardiogram); and, risk of death at 30 days after surgery was estimated at ≥ 15% or greater, and the risk of death or irreversible complications within 30 days was < 50%.

Baseline demographics were comparable between TAVR and SAVR cohorts, but the prevalence of diabetes was significantly higher in the latter group. Surgical risk assessment included consideration of Society of Thoracic Surgeons (STS) predicted Risk of mortality estimate and allowed the heart team to identify other risk factors that were not captured in the STS risk model. A STS predicted risk greater than 10 was considered high, and the average scores in this trial were 7.3 and 7.5 in the TAVR and SAVR groups respectively, and lower than anticipated. Dr. Labinaz offered that the low risk scores reflect inclusion of lower risk patients. Non-STS assessments of co-morbidities, frailty, and disability were also generally similar between groups.

There was a question whether this is the first time for an indication-free trial. Are operators effective at determining risk, or is STS not a valid score to determine risk in this group of patients? Some general discussions were held about the accuracy of the STS score. It was generally agreed that although helpful, there can be important factors of risk for a specific patient that are not accounted for in the STS score.

Primary endpoint was all cause mortality at 1 year. The transcatheter and surgical rates were 14.2% and 19.1% respectively. After 6 months the groups behaved similarly, as "the harm had been paid upfront". A similar trend was observed at 2-years: all-cause mortality showed a TAVR and SAVR rates were approximately 27% and 37% respectively. All stroke incidences between groups were not significantly different, but tended be lower in TAVR as compared to SAVR patients (8.8% and 12.6% respectively; P=0.10). However, the composite endpoint of all-cause mortality or major stroke was significantly lower at 12 months in the TAVR group (16.3% vs. 22.5%). It is important to note that more than half the strokes happened after the first 30 days, and that identifies another area of concern.

When considering 1-year Major Cardiovascular and Cerebrovascular Events (MACCE), the transcatheter group showed an absolute difference of approximately 7% at 12 months compared to the surgical group (20.4% vs 27.3%, respectively; P=0.03). Bleeding, and new onset or worsening atrial fibrillation were roughly doubled with surgery (38.4 % vs. 16.6%; and, 32.7% vs. 15.9% respectively), and acute kidney injury was more than doubled (15.1% vs. 6.0%). Conversely, pacemaker implantation rate was almost doubled with TAVR (22.3% vs. 11.3%).

Echocardiographic Findings:

• TAVI valve gives much larger EAOs (effective orifice area) compared to the surgical valve at all-time points (no sing of early deterioration)
• Big EAO's important when we talk about TAVR and strategies for repeated procedures

Post implant there were significant differences between TAVR and SAVR at each time point for both EOA and mean gradient. TAVR valves gave a much larger effective orifice area (EOA) compared to surgical valves at all-time points to one year (no sign of early deterioration). This raised discussion regarding differenced in valve design that influence EOA. Big EAO is important when we talk about strategies for repeated procedures. The average aortic valve size is 23mm (open surgery), is rigid, and has a cuffed ring which can't be oversized due to risk of coronary obstruction, in order to put in larger surgical valves another procedure such as aortic root enlargement is required. In comparison, the TAVR valve is available in sizes of 26 to 29 mm, and the stent is a thin-walled metal. TAVR valves must be oversized to get a good seal. Also, the CoreValve is oversized much more than SAPIEN (balloon expandable) valves. Incidence of paravalvular leak and regurgitation is always higher with TAVR than with surgery, and this held true in the findings, that showed significant differences at all time-points up to one year. Pacemaker implant rates were also high for the CoreValve group. This rate may reduce over time with improvements in implantation techniques.

Subgroup analysis showed that age, gender, or BMI did not make a difference (TAVR vs SAVR) for 1 year mortality.

The PARTNER 2A Trial

The Partner 2A was a randomized trial that enrolled intermediate risk (operable) patients with STS scores ≥ 4% (Avg. 5.8%). There were 1500 patients in the transfemoral arm, 500 in the non-transfemoral (transapical/transaortic) arm, and the primary end point was all-cause mortality or disabling stroke at 2 years. The TAVR device studied was the SAPIEN XT. Inclusion criteria was severe aortic stenosis (AS) defined as echo-derived AVA ≤ 0.8 cm2 and mean AVG > 40 mm Hg or peak jet velocity > 4.0 m/s; cardiac symptoms defined as NYHA functional class ≥ II; and, intermediate risk which was (i) determined by the multi-disciplinary Heart Team (ii) using a guideline STS ≥ 4% and (iii) adjudicated by a case review committee.

Statistical Analysis Plan:

The primary hypothesis was non-inferiority (ratio of 1.20) of test (SAPIEN XT) versus control (surgery) for all-cause mortality or disabling stroke at 2 years. Assumptions for 1:1 randomization: event rate: 30% in both trial arms, and power at 80%. 2032 patients were randomized with approximately 98% follow-up in both arms.

Emphasis was brought on to the evolution in the devices. The SAPIEN XT was first implanted October 30, 2012. This valve is no longer used in Canada, as practice has moved to using the SAPIEN 3. This is the problem with providing recommendations; by the time the data is available, the end users are already using the next generation device which is generally much improved. For example, the SAPIEN 3 device has a cuff that reduces PV leak rate, and smaller sheath sizes which reduce vascular complication rates.

Patient Characteristics:

Average age of 81.6 years, average STS score of 5.8%, good LV ejection fraction, approximately 20% had moderate to severe MR, frailty was close to half, 10% had permanent pacemaker, and 1/3 had COPD.

Procedural Characteristics:

Anesthesia time, procedure time and ICU / total time in the hospital were significantly less with TAVR. Procedural deaths were equally low in both groups at ~ 1%.

The primary non-inferiority (NI) endpoint (ITT analysis of all-cause mortality or disabling stroke) at 2 years was met. The TAVR and SAVR groups were not significantly different (19.3% vs 21.1%), but TAVR was numerically better (based on combined transfemoral and non-transfemoral access routes). The SAPIEN XT valve met the non-inferiority end-point (relative risk ratio of 1.2) with a significant p-value (0.001). Also, all subgroup analysis favored TAVR expect transthoracic access route, which favored surgery. This indicates that access site plays a role in patient outcomes.

Surgery vs. transfemoral TAVR ITT analysis for all-cause mortality or disabling stroke at 2 years showed a significantly lower rate in the TAVR group (16.8% vs. 20.4%; P=0.05). Other clinical endpoints at 2 years (ITT) showed significantly more life-threatening bleeding with surgery than with TAVR. Major vascular complications were higher with TAVR (using SAPIEN XT not SAPIEN 3 in this trial). New permanent pacemaker with surgery was more than what was predicted and with TAVR was less then what was predicted. The need for a new permanent pacemaker was almost identical between the two options (11.8% TAVR vs 10.3% SAVR). Re-intervention and endocarditis were slightly higher with TAVR compared to surgery.

It was noted that moderate paravalvular regurgitation with TAVR at 2 years was 8.0% while with surgery it was only 0.6%, which is greater than a 10-fold difference; in the S3 trial (SAPIEN 3) moderate valvular leak was only 3%.

Paravalvular leak (PVL) is BAD, and is demonstrated by the following data (all-cause mortality at 2 years): none, trace or mild PVL is associated with mortality rates from 13.5% to 14.1% at 2 years, while mortality rates associated with moderate or severe PVL were significantly increased (34.0%; P<0.001). Therefore, reducing PVL is an important goal which has been demonstrated by the LOTUS and SAPIEN 3 devices that have low PVLs.

1. Consultation before Treatment: For intermediate risk patient populations, is the role of the Heart Team still important and what are the dynamics like here between the surgeons and the interventionalists? How much weight is allocated to patient preference and how is this considered by the various Heart Team members?

   All TAVR programs in Canada have a Heart Team. These teams function differently in different institutions but have an important role in intermediate and high risk patients. However, the definition of intermediate risk is arbitrary and difficult. The STS Score is only a tool and shouldn't be the only mechanism used to determine the definition. A heart team with multidisciplinary expertise is essential in determining if a patient is a good candidate for a specific procedure.

   Will there be a role for the Heart Team when low risk patients are included? Do we have the resources in Canada, to have a surgeon and cardiologist, to scrub into every case? Thought – a gradation of team for intermediate to low risk patients. Perhaps some low risk procedures would only have 1 operator.

   Patient's preference is always considered, and informed consent is essential.

   Valve leaflet immobility and thrombosis: Valve thrombosis and leaflet immobility have been identified as potential issues in up to 40% of patients receiving TAVR's detected using a variety of approaches. This has also been identified in surgical bioprosthetic valves too. The leaflet mobility and thrombus issues can be treated in most patients with antiplatelet/anticoagulant therapy. Should patients receiving TAVR's or surgical bioprosthetic valves be followed more frequently to identify patients at risk for antiplatelet/anticoagulant therapy? Does this concern increase as healthier and younger patients are treated?

   It doesn't matter if you are young or old; if the patient's valve doesn't work, then that's a problem. References were made to "Possible Subclinical leaflet thrombosis in Bioprosthetic Aortic Valves" (Makkar et al.). There are no clinical characteristics to identify who will have leaflet immobility. CT Imaging shows reduced leaflet motion. Clinical outcomes: Pooled registry data suggested that reduced leaflet motion can result in increased neurological events (TIA), but no such relationship was seen in the PORTICO IDE Study data. Subclinical Thrombosis in Bioprosthetic Valves is being studied and there continues to be interest in learning more. At this time, it is early to have specific clinical recommendations.

   Galileo sub-study is ongoing, and investigates high risk patients who were not good candidates for oral anticoagulant (OAC) preventative treatment.

2. Valve Durability: Preclinical bench testing for the durability of bioprosthetic aortic valves including TAVR's is usually stopped at the equivalent of about five years in vivo. Chronic animal studies are usually carried out with small numbers of animals for periods of about six months. It is assumed that the durability of TAVR's and conventional bioprosthetic valves are similar. Differences in design and the need for compression and folding to fit the valves into the delivery systems may result in damage leading to decreased durability. A recent report from Dvir at EuroPCR cautioned that the durability of TAVRs might not be comparable to that of SAVR's. The study followed patients who received early model TAVRs (mostly early balloon expandable Sapien valves). Using a Kaplan-Meier analysis, Dvir reported that the curve for freedom from valve degeneration drops from 94% at 4 years to 82% at 6 years and to approximately 50% at 8 years among surviving patients. Dvir estimated that approximately 40% of degenerated valves would eventually fail. Should TAVR Manufacturers be required to provide long-term premarket and post market data (five to ten years) to demonstrate equivalence to surgical biprosthetic valves?

   Full paper has not yet been published therefore the following information was extrapolated from the abstract. The Long Term TAVR Durability (Dvir) study enrolled 704 patients with a mean age of 82, and includes more than 5 years of follow-up. TAVR procedures performed between April 2002 and May 2011. 378 TAVR patients (Sapien XT 36%, Sapien 50%, Cribier-Edwards 14%) were included. Of those, 100 survived more than 5 years and were examined for moderate to severe intra-valvular AI (not present at 30 day eco) and/or stenosis (mean gradient > 20 mm Hg). 35 cases had valve degeneration (2/3 due to AI and 1/3 due to stenosis). The estimated 8 year rate of structural degeneration is 50%. While ten year data is key, the data aren't there yet as some centers are only coming up on the 10 year mark. However, after four years of experience with CoreValve transcatheter heart valves, none showed deterioration in valve function. In addition, the five year outcome, out of Vancouver, after transcatheter aortic valve implantation, failure rate is very low. The first five years are considered a "free pass", as even poor performing valves have historically performed well for the first five years, but the period from 5 to 10 years is considered critical.

3. Access type: According to many clinical guidelines, alternate access routes for TAVR are associated with worse outcomes. As considerations move to intermediate risk patient populations, is it justified to consider alternate access routes when the candidates may be reasonable surgical candidates. Please discuss data available for this (iliofemoral vs. non-iliofemoral such as subclavian, direct aortic, transapical, etc.)

   A meta-analysis of transcatheter implants was conducted approximately 1 year ago (Chandrasekhar et al. 2015) and compared access types across 17,000 patients including transfemoral, transapical, direct aortic, and subclavian (did not include the new transcaval). At 30 days, there was close to a 50% less mortality by transfemoral versus non-transfemoral routes; however, when compared to subclavian or axillary access, there we no significant difference. Comparison to transapical showed that the transfemoral route wins hands down. 30 day stroke rates were not different between transfemoral versus non-transfemoral routes, but transfemoral was superior to subclavian/axillary routes. Therefore, transfemoral should be the first choice, and the other options should be considered secondary. Transapical is a last resort, and intermediate risk patients would be sent for a surgical valve instead. In some cases, a drop of approximately 10% in LVEF has been observed for patients treated with a transapical approach.

4. Valve in valve: Similar to different access routes, valve-in-valve indications may have limitations when considering intermediate risk patient populations. Is there evidence that valve in valve indications would be acceptable in an intermediate patient population? Should patients with failed TAVR valves be considered candidates for TAVR in TAVR procedures? What evidence is there that Patient/Prosthesis Mismatch can be adequately controlled. What evidence is required to allow you as a clinician to feel comfortable doing this?

   Not a lot has been done on patient prosthetic mismatch (PPM) and TAVR. A German study was conducted on Redo TAVR, and failed TAVR looking at PPM after the second TAVR. The results indicate that PPM does not seem to be a big deal with TAVR in TAVR procedures. The reason for this is that large valves, 29-31mm in size, are being inserted, and the same size valve is put in due to TAVR oversizing, which does not create a smaller orifice in PPM. However, the durability of the valve may pose an issue. If there were a calcified valve, TAVR stent and another stent beside, the stent may cause an issue. PPM occurs more often with SAVR than with TAVR. One study showed the incidence of PPM at 12 months in a TAVR cohort was 7.0%, as compared to 20.7% in the SAVR group (p < 0.0001). If PPM is predicted in a patient, regardless of the risk, going TAVR instead of SAVR would be a better initial strategy, unless surgical root enlargement is a viable option. However, most surgeons will not perform TAVR inside a 21 mm surgical valve, as that size is too small.

   The published study by D. Dvir, "Transcatherer Aortic Valve Implantation in Failed Bioprosthetic Surgical Valves", was discussed. Time to failure is 10 years. Regurgitation alone occurs less frequently as the mechanism of valve failure than stenosis alone, and regurgitation and stenosis combined occur just as frequently as regurgitation alone. Clinical outcomes were considered good, and TAVR with failed valves larger than 21 mm showed medium term outcomes, although more data will be needed. Predictors of 1 year mortality include: valve size (smaller than 21mm is worse); failure type (regurgitation is better than stenosis); transapical access is worse; and as STS score increases, the outcomes worsen.

5. Patient selection: With the possible expansion of indications for use into intermediate risk patient populations, what clinical concerns increase in terms of the patient population that is now represent. Are there any additional concerns as the patient population becomes larger and presumably more diverse? Are there concerns with treating intermediate risk patients that were not of significant concern with high risk patients that had a reduced life expectancy?

   New concerns in treating low to intermediate risk patients include: higher incidence of bicuspid valves (TAVR in these patients may not behave as well); and, maintaining access to coronary arteries (E.g. a 60 y/o patient undergoes SAVR and lives to 85 may develop coronary artery disease that requires intervention). It is important to ensure the stent frame of the valves do not cover the coronary ostium and prevent coronary access. Other considerations include the impact of permanent pacemakers, reduced hospital length of stay/resources, impact on hospital resources/physicians (smaller team), careful screening and importance of heart team, and interaction between treatment of aortic stenosis and other cardiac disease.

   The ensuing discussing centered on findings from a Montreal Heart Institute study that identified a 20% mortality rate or re-operation at 10 years for mechanical valves not performing well. The long term performance of TAVR valves will have to be put in a realistic and appropriate contemporary context.

Dr. Yee left the meeting.

7. Transcatheter Mitral Valve Replacement (TMVR)

Dr. Mark Peterson, Invited Guest Speaker

Cardiac Surgeon in Toronto, ON

There are a host of therapies that may change the treatment of mitral insufficiency and mitral stenosis. Health Canada would be interested in an update on the potential transcatheter mitral valve replacement devices, their recent clinical results, and future directions for these technologies.

1. What are the challenges for the design of transcatheter mitral valves due to the complex anatomy and physiology of heart?
2. Several transcatheter mitral valves are currently being tested in human, such as Tiara, Caisson TMVR, CardiAQ Valve, and the Tendyne Transcatheter Mitral Valve Implantation System. What current clinical evidence is available? What clinical data is expected to be available in a few years?
3. At present, does the clinical evidence point to an appropriate indication for use or specific population?
4. What clinical evidence would be considered to be sufficient for regulatory approval for a specific population? What clinical endpoints should be assessed for the safety and effectiveness of this type of device? How much data is required to move this technology from ITA and SAP in the regulations to licensing?

Dr. Peterson disclosed his affiliations and proceeded to address the questions posed by Health Canada. He started his presentation by discussing TMVR, TAVI and TAVR:

1. TMVR: transcatheter mitral valve replacement
2. TAVI or TAVR: transcatheter aortic valve implantation or replacement

He noted that TAVR is becoming, in some places, the standard of care compared to the Surgical Aortic Valve Replacement (SAVR), which was the golden standard of care in the past. Based on the European Aortic Valve Procedures data the surgical valve implantation and TAVR rates are increasing year after year. In Ontario, TAVR represents 30% of all valve replacements, and demonstrates Canada is following the European trend.

TAVR has been shown, in global randomized trials (PARTNER Trial Cohort A), to be as good as SAVR for high risk patients at 5 years, and with similar rates of all-cause mortality or disabling stroke in intermediate-risk patients at 2 years, as presented by Dr. Labinaz earlier this morning.

TAVR has matured and it is now present in the Acute Cardiac Care Association's (ACCA) European Guidelines, and the current algorithm embraces the heart team concept. The guidelines for the decision making process are followed for intermediate to high risk patients in order to decide whether the patient is a good candidate for TAVR. First-in-man trials were headed by Dr. Cribier in 2002, and mortality reached 30%. However, there has been a large evolution of devices and are now available in a variety of sizes. Also, multiple iterations have resolved various issues including valvular access, sheath size, etc. In 2016, there is only 1 of 11 cardiovascular programs in Ontario that is not performing TAVR; therefore, TAVR has become a commonly accepted technology.

However, there are no published, randomized, Phase I trials for TMVR because they don't exist. What is the reason for this discrepancy? Mitral valve disease is more prevalent compared to aortic valve disease (based on a population based study published in Lancet, 2006), although aortic disease is much more commonly treated. As you age, the prevalence of moderate or severe mitral valve or aortic valve disease increases and there is a significantly larger increase in moderate or severe mitral than aortic valve disease in those aged 55 and older.

Dr. Peterson noted that a large portion of mitral regurgitation (MR) patients are left untreated, either being deemed ineligible for surgical treatment or being denied surgical intervention. However, if you have no mitral regurgitation, your survival rate at 5 years is very good despite having coronary disease. If you have severe regurgitation, your 4 year survival rate is dismal. This is a clear signal – if regurgitation increases, the survival rate decreases.

The structure of the mitral valve is different from that of the aortic valve, with much more complexity. Understanding the mitral valve will help with design and technical challenges with TMVR. Also important is that mitral disease represents a much more heterogeneous population of diseases. Any one component of the mitral valve may be diseased (e.g. leaflets, annulus, ventricles, etc.) which can cause mitral regurgitation, as opposed to the aortic valve disease, in which the vast majority of patients are treated for calcified aortic stenosis.

Mitral disease can be divided into two types: if the valve itself is affected, it's called a primary mitral regurgitation, and can occur from degenerate leaflets (rheumatically or due to endocarditis); and if the disease occurs in the ventricle, it is called secondary or functional mitral regurgitation. In degenerative disease, the cords may be ruptured (Barlow's disease which causes mitral regurgitation). Diseases of the ventricle include dilated cardiomyopathies (in which the mitral valve is pulled into the ventricle, resulting in mitral regurgitation), long standing coronary disease, or geometric shape change in the ventricle (causing the valve to become severely incompetent through tethering or displacing of the apparatus). In conclusion, the heterogeneity of the disease and the complexity of the valve structure are real challenges for engineers who are designing a single fit for all types of valves.

Current treatment options for primary MR include surgical repair, and there are options for the use of robotic or thoracoscopic assists which allow treatment of patients through minimally invasive incisions (the advantages of which are debated). Secondary MR as it relates to coronary disease is treated with medical therapy, or resynchronization therapy.

The question remains, what do we do with the large population of MR patients who have either fallen out of the high risk category, or are ineligible for surgical treatment and denied surgical intervention? How do we determine the proper treatment method for those patients? Of surgical candidates, up to 50% are not referred to surgery, even if a surgical indication exists.

The unmet clinical challenge is addressing those patients not treated. Surgical treatment rate of moderate to severe patients with degenerative MR (DMR) is ~ 53%, and for those with functional MR (FMR) is ~16%. Low treatment rates in FMR patients are the result of previous guideline that didn't stress surgical intervention, and low treatment rates in DMR are due to patients being asymptomatic, with stable LVEF, and substantial comorbidities contributing higher risk.

Percutaneous options for mitral valve disease were described.

1. Patients with rheumatic mitral stenosis, provided they have an anatomically suitable valve, would be candidates for balloon valvotomy.
2. For patients with congenital regurgitation, but who are too high risk for open surgery, the mitral clip is the dominant repair strategy.
3. Valve in valve therapy can be done by implanting a TAVR valve. The rigid sewing ring of the mitral valve functions as an anchor. In this case you are not treating the mitral valve anatomy but treating the bioprosthetic valve, and this approach seems to work well.
4. Percutaneous closure devices to treat peri-valvular mitral regurgitation
5. Devices with specific TAVR-like design for mitral valve diseases

Why we need alternative care rather than surgery?

A case study of transcatheter mitral valve replacement (transvenous-transseptal mitral valve-in-valve) was presented:

72 year old male patient with six months of progressive shortness of breath (SOB), who is now NYHA Class III. Patient's medical history (PMHx) includes: Hancock II bioprosthetic AVR and MVR in 1998; atrial fibrillation (AF), sick sinus system (SSS) and on warfarin; coronary artery disease (CAD) with previous non-ST elevated myocardial infarction (NSTEMI) with percutaneous coronary intervention (PCI) in the circumflex coronary artery (Cx) in 2013; and, smoker with severe chronic obstructive pulmonary disease (COPD).

Echo showed dilated left ventricle, preserved ejection fraction, severe MR due to flail leaflet, and valve gradients of 8-9 mmHg. Catheterization showed patent circumflex stent. Right heart catheterization (RHC) showed systolic, diastolic, and mean pulmonary artery pressures of 53/20/39 mmHg respectively. Pulmonary capillary wedge pressure (PCWP) was 23 mmHg, and the V waves were 45 mmHg. CT scan showed the mitral valve annular area was 5.09 cm2, and was considered low risk for left ventricular output tract obstruction (LVOTO).

In consultation with the heart team (key players: surgeons, cardiologist, fellows (surgical and interventional) etc.), all data is exhaustively vetted. The discussion of options would have included: redo open mitral valve replacement (leading to replacement of the aortic valve), or either trans-apical, or trans-septal valve-in-valve.

The actual outcome:

Cannulated the femoral vein, went up across the right atrium, and through the atrial septum (transeptal puncture). The challenge was determining how to get the prosthesis to the mitral space. Options were transapical or transeptal delivery. However the turn via the transeptal delivery does not allow for much freedom of motion – both methods provide anatomical and technical challenges. Valve-in-valve is off label but many surgeons believe there is a solid indication and niche in performing this technique in all bio prosthetic valves. The anatomical constraints of the mitral valve are negated by the bio prosthetic valve.

The problem is getting the device where it needs to be. The device doesn't go through the septum therefore you are in the atrium, and then you must proceed through the septum then into the bishop. It was stated that the valve may not need to be modified but a better delivery system.

In the early days the mitral clip had numerous complexities in trying to position the device. Currently, with experience and over time the complexities are less. However, the design idea for the TMVR was that for the valve would be easier to implant than a mitral clip although getting there is still a significant limitation. As opposed to the mitral clip, once you have the properly designed TMVR in place you eliminate, not reduce but eliminate the mitral regurgitation (MR) and then there less recurrent MR.

Design features that are different then a TAVI valve:

• Atrial flange that is designed to sit on top of left atrium, on top of the mitral valve. Note that the mitral valve is not symmetrical but rather D-shaped. This was taken into account when they designed the flexible strut to sit closer to the aortic valve.
• One size fits all valve with a 29 mm cylinder.
• Coating was designed to create seal and contribute to healing, however it may contributed to some of the thrombosis.
• Unique to the device as the paddles, which are designed to open up and as the valve is centered in the middle of the ventricle, the paddles are closed, one paddle captures the interior of the leaflet and one paddle captures the posterior leaflet.

One similarity is the leaflet inside the cylinder are bovine pericardial.

Fortis Valve was acquired under Special Access Program (SAP) and under compassionate use. The fifth case was conducted at St. Michael's hospital in Toronto. 7 of 11 patients survived.

There was a discussion among the members around tethering and the necessity of the cords of the device.

Lessons learned are similar to the TAVI story.

Factors that effected the outcomes were both technical and patient selection, in particular, the recognition that cohort c patients (those with a STS score greater than 15%) probably don't benefit don't benefit from TAVR however if you select cohort c patients for TMVR or patients with ejection fractions of 10-15%, they would also do poorly. Other lessons learnt were the anatomical issues relating to anti-coagulation.

Where is TMVR development today?

In the last two years, industrial sector saw $2 billion being spent which resulted large medical devices companies acquire nascent technology in this filed and approximately 33 different devices are various iterations of devices being developed. The USA has recognized that in the TAVI field they were behind Germany, Europe and slightly behind Canada and were only to put TAVI valves in special circumstances. Therefore they developed an early feasibility pathway that would allow them adopt and support innovation and new technology. Currently there are 6 devices approved for early feasibility in the USA, two of them are the CardiaAQ device and the Medtronic Intrepid device. All devices are various forms of purpose built transcatheter mitral valves to be deployed either transapically or transeptically.

More commonly used devices are:

1. CardiaAQ was designed for both transeptal and transapical delivery, has an expanding nitinol frame, it preserves the cords and utilizes the leaflets, open frame promotes atrial flow, intra-annular skirt to minimize paravalvular leaks and to minimize LVL obustruction.
2. Medtronic Intrepid is similar in design to the CardiAQ, it was an inner and outer stent for fixation and sealing with a 27 mm bovine pericardial valve and is effective in relieving mitral regurgitation.

Challenges for TMVR:

• Patient selection and screening (example: hard to justify applying this technique for patients who are candidates for mitral clips, anatomical criteria and screening results)

Discussion around being a candidate for other procedures that may be standard of care versus undergoing a TMVR. For example, once you have a mitral clip implanted you are no longer eligible for TMVR in the future.

In conclusion, transcatheter mitral valve replacement is complex, certainly anatomically more complex than TAVR, more patients with heterogeneous group of diseases, group of feasibility in first in man studies, also users to highlight the anatomical and related factors that have led to the successes by continued device refinement, patient selection and screening it will lead to a viable therapy to treat TMVR.

TMVR First-in-man Clinical Experience: Fortis (Edwards)

Background:

Mitral regurgitation is more common than aortic stenosis in the elderly population, who are also at higher risk, and survival is poor if left untreated. Mitral surgery risk increases with: advanced age; multiple comorbidities; prior CABG; and low ejection fraction. Less invasive mitral valve therapy has the potential to treat high risk patients. All cases of human implantation performed on "compassionate" grounds (not within trial protocol). 11 patients were implanted worldwide, survival is 7/11 (64%) to date. Causes of death were acute kidney injury (AKI) and multiple organ failure (MOF) (4 days), heart failure (76 days), valve thrombosis (15 days), and incomplete P2 capture with conversion to mitral valve replacement/sepsis (7 days).

Lessons learned from the Fortis experience: it was not dissimilar to the TAVR story. The Fortis was too thrombogenic. Outcomes were largely influenced by degree of LV dysfunction, pulmonary disease, renal disease, redo surgery, etc. The mitral patient may require longer to see benefit than the aortic. Mitral valve pathology related to annular dimensions must be considered. The anatomy of P2 and pre-operative imaging must exclude potential for LVOT obstruction. Saddle shaped valve annulus may increase challenges in reducing paravalvular leakage.

Currently 33 TMVR devices in development, 6 have been approved for early feasibility studies in US (CardiaQ, Tendyne, Neovasc, Intrepid, Caisson, and Mvalve/Lotus).

CardiAQ – Edwards TMVR

The first in human trial ran in June of 2012. The valve can be implanted via a transseptal or transapical route. Has a unique anchoring mechanism that utilizes native leaflets and preserved chords. To date 13 patients have been treated with 2 procedure-related and 7 non-valve-related deaths. The RELIEF Trial is set to start.

Intrepid – Medtronic TMVR

The outer stent is designed for fixation and sealing, and the inner stent houses a 27 mm bovine pericardial valve. It also houses a flexible brim which aids imaging during implantation. To date, 27 implantations have been performed with 4 deaths related to the procedure, none related to the device, and 3 not related to the procedure or the device. Pre-procedure MR evaluation showed 19 patients with MR of 4+ and 8 with MR of 3+. The latest follow-up saw 22 patients with MR of 0+ and 2 with MR of 1+.

In conclusion Dr. Peterson summarized that TMVR is very complex, but 'Proof of Principal' is established by first-in-man and early feasibility studies. Ongoing work will establish better patient selection, procedural steps, and post-operative management, which includes anticoagulation. As a note on patient selection, it was highlighted that TMVR may not perform well in patients with a very low LVEF.

8. Drug Coated Balloons (cardiac)

Dr. John Ducas, Committee member

Dr. Kong Teng Tan, Guest Speaker

OBJECTIVE: Assisted by recommendations given by SAC-MDUCS, Health Canada has licensed Drug Coated Balloons (DCBs) for both peripheral and coronary use based on superior effectiveness compared to POBA in terms of endpoints such as clinically driven revascularization and patency. As the technology becomes more readily available in Canada, it would be helpful to understand how it is being used in clinical practice in Canada, whether there are any emerging concerns associated with its use, and whether POBA as a comparator is still appropriate.

Currently, there are several DCBs licensed in Canada. The Pantera Lux is indicated for BMS-ISR in the coronary arteries. The Lutonix 035 DCB PTA Catheter is indicated for percutaneous transluminal angioplasty, after pre-dilatation, of obstructive de novo or non-stented restenotic lesions in native superficial femoral or popliteal arteries, with reference vessel diameters of

4 – 7 mm. The Lutonix 014 Drug Coated PTA Dilatation Catheter is intended for percutaneous transluminal angioplasty of obstructive de novo or non-stented restenotic lesions in the native popliteal, tibial, and peroneal arteries ≥2.0 and ≤4.0 mm in diameter.

1. How are DCB currently being used in Canada?
2. What level of clinical data is required to authorize a new license for a new DCB? Is POBA still the best comparator group for DCB studies for a superiority RCT? Are there differences in anatomical locations - coronary vs peripheral?
3. Are there any safety concerns associated with DCBs, especially given some recent clinical trial findings? Studies of interest might include In.Pact Admiral RCT. How long should DCBs be followed to determine their safety profile compared to other interventions?
4. What level of data is required for an expansion of indications for use?
• De novo versus restenotic versus in-stent restenosis?
• Above the knee versus below the knee?
• Longer length balloons and longer length lesions.
5. Specific to the treatment of ISR, discuss the following:
1. Should the standard of care for ISR be PTA/PTCA or DES as the comparator against which DCB should be compared?
2. Are BMS ISR treated differently than DES ISR?
3. Are their differences in DES in the treatment of ISR, e.g. EES vs. other DES
6. Discuss the future directions with DCB in the coronary anatomy and in the peripheral anatomy.

Dr. Ducas disclosed his affiliations and proceeded to address the questions posed by Health Canada. He focused on the use of drug-coated balloons to treat coronary artery disease

He noted that currently the use of drug eluting stent (DES) is the treatment of choice for most PCI because: the anti-proliferative coating drastically reduces neo intimal hyperplasia, devices are highly deliverable, and they have relatively low thrombogenicity, low rates of target vessel revascularization (TVR) rates, and current generation DESs are safer and more effective than first-generation devices. Their use is supported by robust data and experience in multiple clinical scenarios involving complex diseases, involvement of the left main (LM) coronary artery, bifurcations, and S-T segment elevated myocardial infarction (STEMI).

However, there are some known limitations due to the metal cage left behind:

• vasomotility may be affected
• strut malapposition that may lead to late thrombosis
• "jailing" of side braches that may affect future PCI
• full metal jacket that may preclude coronary artery bypass graft (CABG) procedures
• complex bifurcation lesions remain a problem
• applying multiple layers of stents is required in certain patients with in stent restenosis

The use of drug-coated balloons (DCBs) represents another treatment option. A matrix coating sits on the balloon surface, and contains an excipient and active drug (E.g. Paclitaxel) with a high retention rate of up to one week, or longer. Inflation of the balloon deposits the highly lipophilic drug onto the surrounding tissue surfaces. First generation balloons had problems with control of drug delivery that included loss of drug substance during delivery, non-uniformity of drug coating on the balloon, embolization of the drug, and increased serum drug levels resulting from inflation of multiple balloons.

Dr. Ducas highlighted treatment gaps that have been potentially addressed by using DCBs by referring to a past SAC-MDUCS discussion (2010-12-10):

• In-stent restenosis (ISR)
• Small vessels
• Bifurcations
• Acute myocardial infarctions (AMI)
• Chronic total occlusions (CTOs)

Next he referred to the Health Canada, Regulatory decision summary for the PANTERA LUX PACITAXEL RELEASING PTCA BALLOON CATHETER that indicates the application was filed 2012-08-07, and a regulatory decision was made on 2015-07-09. The manufacturer cited the superiority of paclitaxel eluting balloon (SeQuent Please) to POBA, and then invoked the the similarity of the Pantera LUX to the SeQuent Please to justify approval of the Pantera LUX. At that time the SAC-MDUCS committee endorsed POBA as an appropriate comparator.

Dr. Ducas noted that recently there have been number of technical improvements made in DCB technology including:

• advanced folding technologies to minimize drug loss during delivery
• altering drug concentration
• altering size drug particles
• selection of non-crystalline formulations
• drug encapsulation using nanoparticles allowing incorporation sirolimus on DCB
• Use of zotarolimus (more lipophilic)

Several sirolimus-DEB already received CE-mark approval.

However, clinical experience very limited. Dr. Ducas noted that while DCBs have proven efficacy, clinical experience very limited. The data are limited to small studies with first generation DCBs, and the class effect is unproven and there are no randomized head to head trials of DCBs.

Number of studies were presented next.

A Swedish registry –SCAAR (Per Bondesson et al. 2012) that included 1129 patients, compared two DCB. The cumulative risk of restenosis was both greater and earlier (time since PCI) with ELUTAX as compared to the SeQuent Please at one year follow up.

In a comparison of the efficacy of the Xience everolimus eluting stent and SeQuent drug eluting balloon in treating ISR, freedom from MACE (cardiac death, MI, and TVR) was significantly greater in the Xience group (91% vs 85%; P<0.04;n=498) (Alfonso et al. 2016).

Stella et al. (2012) describe a multi-center randomized comparison of DCB plus BMS versus BMS versus DES in bifurcation lesions treated with a single-stenting technique: six month angiographic and 12 month clinical results. Pretreatment of both the main branch and side branch with the DCB failed to show angiographic and clinical superiority over conventional BMS, using a provisional T-stenting technique. DES showed superior angiographic results as compared to DCB and BMS.

Dr. Ducas concluded his presentation indicating that DCB was clearly superior to POBA for ISR, but that it was less clear whether DCB is equivalent to current DES devices for any indication. Also, based on a very small, non-RCT study by Kawamoto et al. (2015), the superiority of DCB to DES was uncertain in recurrent ISR with already present multiple layers of. Also, Dr. Ducas emphasized, DCB is not approved for other PCI uses, as limited published data in small vessels and bifurcations are not compelling. He suggested that DCB use in Canada is probably limited (currently < 1% of patients receive DCB), but future innovations may expand their through the Special Access Program.

Dr. Yee returned to the meeting.

Dr. Mitchell left the meeting.

Dr. Tan declared his affiliations. His presentation focused on the use of drug-coated balloons for peripheral vascular disease. He defined some specific terminologies to aid interpretation of his presentation.

Binary restenosis: > 50% restenosis of treated segment that is measured by duplex or angiogram.

Target lesion revascularization (TLR): a clinically driven re-intervention of a previously treated segment, that is usually about half of the binary stenosis rate.

Dr. Tan explained that current endovascular techniques involving stents or balloons are immediately effective at revascularization, but within 6 months patients may suffer multi-level restenosis, and they are frequently more occluded than before the procedure. Plain Old Balloon Angioplasty (POBA) has been shown to have very poor primary patency with lesions greater than 5cm in length. A one month restenosis rate of 23% has been described (Romiti, 2008). Therefore patients with long lesions would require multiple interventions to keep the arteries open, and are associated with worse outcomes.

He indicated that intimal hyperplasia occurs in 20-50% of intervened arteries, but anti-proliferative agents can decrease intimal hyperplasia. Paclitaxel and the "limus family" of drugs have demonstrated this effect, as evidenced by coronary restenosis rates of 20% and 5% associated with bare metal stents and drug eluting stents respectively. He also indicated that wound healing depends on patent arteries. He noted that approximately 30% of patients with critical limb ischemia (CLI) will require interventions following initial endovascular therapy due to restenosis of target lesions, which will occur within 1 to 2 years (BASIL Trial, 2005: Haider, 2006; Giles, 2008; Conrad, 2009). He indicated that DES worked relatively well, but is not used much as people don't like implanting stents in the legs.

The currently approved DES for peripheral vascular disease (PVD) is the ZILVER PTX (Cook Medical). The ZILVER PTX randomized clinical trial compared the Zilver PTX (paclitaxel) with POBA. Primary patency (defined as peak systolic velocity ratio, PSVR < 2.0) rate was significantly greater in patients treated with Zilver PTX as compared to optimal PTA, and PTA at 12 months with rates of 83.1%, 65.3%, and 32.8% respectively. Also, the Zilver PTX was compared to the bare Zilver to determine if the significantly higher patency rate was due to the drug effect. 12 month patency (defined as PSVR < 2.0) was statistically higher in the Zilver PTX group as compared with the bare Zilver (89.9% vs. 73.0%). The reduction in 12 month restenosis rate was 63% (27.0% with the bare Zilver as compared to 10.1% with the Zilver PTX). At 30 months the primary patency remains statistically significant, and at 24 months there was very little target lesion revascularization (TLR) (Dake et al. 2013).

As part of the MAJESTIC trial, Boston Scientific's ELUVIA drug eluting SFA stent coated with paclitaxel provided a primary patency rate of 96.4% (Müller-Hülsbeck et al. 2016), but there were concerns with strut fracture. The ELUVIA coating is similar to that used on the TAXUS DES (coronary stent) with slower release of paclitaxel.

However, despite the positive results, stents are not widely used for fear of stent fractures.

There is lots of movement in legs, hips and knees, and this movement can cause fractures which in turn cause internal hyperplasia, and it is usually quite severe.

Dr. Tan discussed the advantages of drug coated balloons as compared to stents and included the absence of stent-related complications, more uniform drug delivery, use in all locations (as well as across joints or major arterial branches), and use in small vessels (stents aren't placed in small vessels because they thrombus right away).

He described the use of DCBs across the knee, and indicated that about 20 DCBs are available in Europe.

DCB design was described as including an angioplasty balloon, coated and bound with a carrier / excipient (urea, sorbitol, BTHC, shellac, isopropamide, citrate ester, polyethylene glycol), and a drug (in Canada all balloons are paclitaxel based). He indicated that drug dose generally ranges between 2 to 3 micrograms/mm2, but is dependent on the excipient and how much drug can be loaded. Also, he noted that coating processes are highly variable (dependent on the manufacturer), and that all design factors will affect balloon efficacy.

He noted that Paclitaxel inhibits cell division by interfering with microtubule function, and the more paclitaxel applied, the less neointimal hyperplasia is seen. Balloon drug losses occur during delivery (20%, from handling outside of patient), are delivered to the vessel wall (20%), and wash into the blood stream (40%). 20% of the loaded drug remains on the balloon.

Dr. Tan suggested drug loss may be decreased through coating/excipient formulations that include a higher drug load, ensure uniformity of coating, stability during delivery in blood vessels, uptake of drug to vessel wall (enhanced transfer efficiency), drug retention in vessel, reduced downstream loss (risk of embolization) and balloon design.

Above the Knee

Number of randomized trials exist that compare POBA with DCB for use in the superficial femoral artery (SFA):

InPact SFA (Medtronic), Levant 2 (Bard/Lutonic), and Biolux (Biotronic). All three coated balloons showed higher primary patency rates, and lower clinical TLR rates compared to POBAs. Based on this, Dr. Tan indicated that DCB use was good for SFA applications.

The randomized THUNDER trial investigated the effectiveness of a paclitaxel-coated balloon (PCB) for restenosis prevention in the femoropopliteal arteries at five years (Tepe et al. 2015). Rates of first TLR at five years in patients treated with POBA were significantly higher than in those treated with PCB (55.6% vs 20.8% respectively). Time to TLR was also significantly higher in patients treated with DCB (607 vs. 206 days). Similarly, binary restenosis at five years was significantly higher in the POBA cohort vs. PCB (54% vs. 17% respectively).

Another problem is in-stent restenosis, as it occurs in 30% to 50% of SFA stenting. Available, but ineffective treatments include angioplasty with POBA, cutting balloon, cryoplasty, atherectomy, relining with a bare stent, the use of drug-eluting stents, or angioplasty with DCBs.

Zeller et al. (2013) showed that 12 month primary patency of femoropopliteal ISR with paclitaxel eluting stents was 79%, based on data from the Zilver PTX single arm prospective study. Krankenberg et al. (2015) showed freedom from TLR at 6 and 12 months was significantly greater in DCB treated patients as compared to those treated with POBA for in-stent restenosis in the femoral artery (90.8% vs. 52.6% at 12 months respectively). Recurrent binary in-stent restenosis at six months was also significantly lower in the DCB vs. POBA cohorts (15.4% vs. 44.7% respectively).

Treatment for instant-restenosis

1. PTA
2. Cutting balloon
3. Cryoplasty
4. Atherectomy
5. Recline with bare stent
6. Recline with covered stent
7. DES

However, options 1 – 5 have all proven to be in-effective.

Below the Knee

When treating below the knee using coronary DES off-label for infrapopliteal disease, consider that DES has lower target lesion revascularization rates, the cost of DES is much higher than POBA, DES is currently used off-label for failed POBA, and multiple stents are often needed as the longest coronary stent is only 3.8 cm.

Randomized trials comparing DES (sirolimus and everolimus coated) to BMS for infrapopliteal lesions < 4cm have shown 12 month primary patency rates of 71% to 85%, and 54% to 56% respectively.

Schmidt et al. (2011) reported on the treatment of long infrapopliteal lesions (176 ± 88 mm) using a paclitaxel-coated balloon (InPact Amphirion, Medtronic); at 3 months the restenosis rate was 27%, as compared to a historical POBA control of 69%.

The DEBATE-BTK randomized, single-center trial compared DCB vs. POBA treatment for long below the knee de novo lesions in diabetics. Freedom from TLR rate was significantly higher following DCB treatment, as compared to PTA at one year. Binary restenosis rate was also significantly less in the DCB group (27% vs. 74.3% respectively) (Liistro et al. 2013).

In a multicenter, randomized comparison of 12 month efficacy endpoints for DCB vs POBA for the treatment of critical limb ischemia in infrapopliteal arteries, Zeller et al. (2014) showed that the InPact Amphirion DEB exhibited comparable efficacy to POBA with respect to late lumen loss and binary restenosis rates, but there was a trend towards increased major amputation rate through to 12 months with the DCB (Zeller et al. 2014).

Similarly, the 12 month results from the BIOLUX P-II randomized trial involving Biotronik's Passeo-18 LUX DEB showed that TLR rates, patency loss, and target extremity amputation were comparable to POBA.

Outcome differences between above and below the knee interventions are attributed to: smaller arteries, longer lesions, suboptimal drug delivery, calcifications and embolization of coating material [below the knee].

Questions:

1. How are DCB currently being used in Canada?

   Both above the knee (SFA and popliteal) and below the knee uses are largely second line of treatment after failure of conventional angioplasty or conventional balloon/stent (purely because of cost). Their use as the primary device only occurs in selected centers. Their use as a below the knee primary treatment is rare. Used in patients with failed previous angioplasty.

2. What level of clinical data is required to authorize a new license for a new DCB? Is POBA still the best comparator group for DCB studies for superiority RCT? Are there differences in anatomical locations - coronary vs peripheral?

   DCBs are designed differently hence RCT is required for all DCBs regardless of locations (above or below the knee). In addition, coronary artery disease is not similar to peripheral vascular disease, particularly because of forces acting on limb arteries (radial compression, longitudinal compression/extension, flexion and torsion) as compared to coronary arteries. In Germany, they approved all drug coated balloons for SFA and popliteal arteries.

3. Are there any safety concerns associated with DCBs, especially given some recent clinical trial findings?

   For SFA and popliteal arteries, the risks are low as shown in current clinical data (many different trials with various manufacturers all showing consistent results). However, there are concerns with below the knee interventions; two randomized trials have shown poor results in infra-popliteal segment, and there is a major concern of worse outcome with DCB as compared to POBA. DCB in renal/mesenteric arteries and leg bypass –insufficient data for any recommendation. For dialysis interventions, randomized trial shows slight improvement in patency compared to plain balloon

4. How long should DCBs be followed to determine their safety profile compared to other interventions?

   Every balloon manufacturer has to be assessed individually. For above the knee applications, 1-2 year data should be sufficient to determine if there is a positive outcome. For below the knee, 6-12 month data should be sufficient to determine if there is a positive outcome.

5. What level of data is required for an expansion of indications for use?
   1. De novo versus restenotic versus in-stent restenosis?

      This issue was not specifically addressed except as noted below.

   2. Above the knee versus below the knee?

      For above the knee, the data is conclusive that DCB is superior to POBA whether it is a de-novo or restenotic lesion. Below the knee more data are needed.

   3. Longer length balloons and longer length lesions.

      Longer lesions are associated with poorer outcomes regardless of treatment type. With longer balloons there is a limitation, in that the risk of embolization of balloon coating material increases.

6. Specific to the treatment of ISR, discuss the following:
   1. Should the standard of care for ISR be PTA or DES as the comparator against which DCB should be compared?

      ISR should be treated by DCB as primary treatment. Placing a DES is indicated if there is stent fracture or dissection proximal or distal to the pre-existing stent.

   2. Are BMS ISR treated differently than DES ISR?

      Cardiology literature shows that DCB and DES are superior to POBA alone (ISAR-DESIRE 3). BMS ISR is generally longer, diffuse and occurs earlier (less than 12 months). DES ISR is focal, often at both ends of the stent and occurs later (after 12 months). DES restenosis is generally shorter and less severe than restenosis observed following POBA. DCB and DES have both been used with good success, but generally stenting is avoided in the small peripheral vessels if possible.

7. Discuss the future directions with DCB in the coronary anatomy and in the peripheral anatomy.

   In Peripheral Vascular Disease, DCB is the first line of treatment for SFA/popliteal disease in many European countries and the USA, with exceptions in severely calcified lesions, where there may be suboptimal drug transfer. In these cases, DES is considered. For below the knee applications, POBA is still the standard of care.

This concluded Dr. Tan's presentation.

9. Pediatric Devices

Dr. Joaquim Miro, Committee Member

Dr. Miro disclosed his affiliations and proceeded to address the questions posed by Health Canada.

OBJECTIVES: Pediatric devices come with specific challenges for the Medical Devices Bureau. Based on the small number of patients that need to be treated with many of these devices, licensing and specific pediatric indications are often not pursued. Regulatory decisions are often made either through the Special Access Program or more rarely through Clinical Trials. To better understand the challenges and use of pediatric devices, we would like to have Dr. Miro discuss these issues and for him to highlight to Health Canada, issues he feels would help Health Canada better regulate these kinds of devices so that appropriate devices and data might be available to clinicians and to Canadians.

Health Canada is interested in finding out more about the use of pediatric medical devices in Canada and some of the specific challenges faced by these devices. Given their limited use and the limited information available about their use in a pediatric population, as well as the challenges associated with running clinical trials in a pediatric population, we are asking for an overview of these devices so that Health Canada can make informed and appropriate regulatory decisions for these devices. Although Health Canada welcomes information on any number of devices associated with treating pediatric patients, areas that might be considered include:

1. An overview of devices and alternatives used for patients with congenital heart disease and some of the unique requirements this may present.

   Dr. Miro noted that pediatric congenital cardiology deals with a very different substrate than adult's cardiology. The affected population is much smaller, and the incidence of congenital heart disease is between 3-5 patients per 1,000 live births. Acquired cardiac disease that includes coronary artery disease, aortic valve disease, mitral valve disease, and atrial fibrillation, is expected to occur in 50% to 100% of individuals. From this, one could predict that everyone will need some type of treatment.

   There is greater lesion variety and complexity in the pediatric population, and patients can be large as well as being much smaller and range from 100 kg to premature infants at 800 g. They have smaller vessels and their anatomy is expected to change as patients grow. This means malformations will change over time and require implantation of different devices because devices will break and not grow with patients, particularly given life expectancies greater than 80 years.

   The small number of patients with specifics needs creates numerous challenges for device development and availability:

   • High cost of research (E.g. bench test studies, animal studies).
   • Lengthy and increasingly burdensome regulatory process (2-3 times the cost and time requirement increases over the past 20 years;
   • By the time the device is approved, there may already be another iteration since technology is evolving rapidly.
   • Interest from industry to develop new pediatric devices is decreasing (must consider multiple complex shapes and sizes to treat low incidence malformations).
   • Industry is more interested in developing devices for adults, including devices for TAVR, since there is more money in the number of devices being implanted.

   Dr. Miro suggested that the complexity and heterogeneity of congenital patients creates a powerful stimulus for innovation.

   He presented an overview an overview of devices and alternatives used for patients with congenital heart disease and some of the unique requirements this may present. He noted that from a regulatory perspective, devices used in congenital heart disease can be obtained through the following processes:

   1. Regular licensing – IFU compliance
   2. Regular licensing – Off-label use
   3. Unlicensed Device: Special Access
   4. Custom made devices: Special access
   5. Regular licensing – IFU compliance
   1. Regular licensing – IFU compliance

      This is the preferred route as most patients, and their parents prefer a licensed device that is designed to treat common malformations. He listed examples of such devices:

      • Septal occlusion devices include: Amplatzer ASD, PFO, and muscular VSD, Gore ASD and PFO.
      • Vascular occlusion devices include: Amplatzer PDA (ADO I, ADO II, ADO II AS), Amplatzer vascular plugs (1, 2, 3, 4), and Coils.
      • Stents and stent grafts include: Palmaz, CP stent, Andrastent XXL for coarctation and pulmonary artery.
      • Balloons include: Tyshak and Nucleus for valve dilation, Z-Med for large vessel dilation, BiB for stent placement, and Mullins for stent re-dilation.

      He described the unique clinical requirements for pediatric devices include: immediate efficacy and safety, long term efficacy and integrity, and miniaturisation of devices (E.g. vascular access delivery systems). He noted that issues with mechanical integrity should be expected, but are acceptable if function of device is preserved.

      Dr. Miro suggested that because pediatric patient populations are small, and regulators require animal and bench testing, perhaps regulators would consider accepting studies and regulatory decisions from other countries in lieu. Furthermore, he suggested the Special Access Program could be accepted as a "path to regulation" in which Canadian investigators could pool clinical results. Creation of registries was suggested as well.

   2. Regular licensing – Off-label Use

      Congenital intervention has a long tradition of adaptation and innovation. Dr. Miro noted that off-label use is more the rule than the exception, as devices are borrowed from interventional radiology or adult cardiology (e.g. coronary, biliary and iliac stents in coarctations and pulmonary stenosis of small patients; coronary stents to keep PDA open (palliation of cyanotic malformations; Sapien valve in pulmonary position).

      Devices are borrowed from pediatric intervention (E.g. Amplatzer Duct Occluder II (ADOII) device made for Patent Duct Arteriosus (PDA's) which isn't that good for PDAs, but is frequently used for VSD); use of several CP stents, designed for the aorta, in the RV, and it's important to remember that the RV is a muscle and that any stent developed for vascular indications is an excellent example of an off-label use with possible consequences (material fatigue, leading to fractures is expected).. Therefore, two or three stents are inserted.

      Dr. Miro suggested the following regulatory considerations. He noted that the initial licensing is where we need to put the efforts, and once a device is licensed it is the duty of physicians to find new applications, discuss with patients, report all results to the medical community, and report dramatic complications to regulatory agencies. He explained that trying to regulate off-label use would be counter-productive for the patients, inefficient, costly, and could pose legal problems (E.g. any physician sued for an off-label complication would try to prove that it was actually allowed, and implicate Health Canada). Voluntary changes in IFU by the manufacturer, to include new established indications should be welcomed. Warnings for adverse events should be updated and included in the IFU.

      He stressed the importance of informing the patient and/or family, when a device is used off label and for the physician to then report on the results whether good or bad.

      A discussion regarding off-label use and concerns for the use of devices in an unintended and different environment ensued. The example used was biliary stents. Bio-compatibility of these devices is reviewed by Health Canada, however their intended use in the biliary and not vascular system, therefore hemo-compatability will not be reviewed or established. As a result, Health Canada cannot provide comments on off-label use in an environment different from the one that is indicated.

   3. Unlicensed devices -Special Access Program (SAP)

      Of developed countries in 2016, Canada is probably the least difficult country in which congenital cardiac devices can be obtained via special access. There is a fine balance between access and scrutiny. Scrutiny is necessary to protect patients against clearly unsafe devices, and to avoid the use of SAP as a permanent business plan that excludes eventual licensing. Many cardiac malformations need to be considered as "orphan diseases" with too few patients to support the effort/investment of regular licensing. Dr. Miro suggested that access to SAP devices gives a definite advantage to Canadian patients, as compared to American patients. (E.g. Amplatzer for ASD-PDA and VSD occluders; bare and covered CP stents; Amplatzer membranous VSD occluders; Gore large ASD occluder; Lifetech memb VSD device (in process); Venus pulmonary valve (in process)).

      Special access will always remain a necessity for cardiac congenital disease, as its disappearance would lead to unnecessary open heart surgeries, and even deaths. It was suggested that whenever possible, SAP should lead to licensing, and perhaps through a special licensing process for devices that treat rare conditions. Also, the relation of trust between Health Canada and physicians will always be primordial; the burden of proving the necessity of a device belongs to the physician. Avoid using special access as a permanent business plan excluding eventual licensing.

   4. Custom made devices - SAP

      In rare instances, no devices are available, including off-label use. The congenital interventionist must find a manufacturer ready to craft the device. Most custom-made devices are size modified adult devices, and fundamental changes in mechanical properties are not expected.

      Dr. Miro suggested Health Canada should be vigorous in vetting these devices for safety for immediate use with less focus on long term use. In pediatric cases, immediate safety and efficacy is considered to be 30 days, while long term safety and efficacy data is often not available; it would take too long to generate this data. When looking at intra-cardiac devices (big devices in small hearts), they are not always expected to last mechanically for life, since in many instances they will be explanted at a subsequent open heart surgery. If they last 2-3 years and break, without directly harmful consequences other than accelerating a surgery that was planned anyway, this may be an acceptable outcome. Metal fatigue should not be the basis of refusal unless the device fails within the first 3 years.

      He also suggested that Health Canada should not vigorously scrutinize a smaller version of a device or delivery systems for which safety and efficacy have been demonstrated for the larger sizes.

      He also noted that large animal testing is not possible with pediatric devices, and that the most Health Canada will see are studies involving 10-12 animals. He suggested that asking for specific studies in Canada for every single device is not realistic, and that Health Canada should accept data from good studies, and licensing and approval decisions from other jurisdictions with good regulatory agencies.

      To gain clinical data, he suggested that Health Canada apply some pressure on applicants to report their results, if the requested device is made available through the special access program (SAP).

      He also suggested that Health Canada to host a session at the Canadian Cardiology Society (CCS) and present on the Special Access Programme. This may include information on pediatric devices approved and used through the SAP in the past year. He indicated this would encourage physicians to share their experiences and information, and thereby help physicians better identify appropriate indications for use of pediatric devices.

      Dr. Miro also proposed the creation of a SAP registry, and reiterated that physicians who access devices through the SAP should be obliged [required] to provide patients' information, devices inserted into patients, and any catastrophic events. Making this information readily available would allow physicians to access vital information about the safety of specific devices/procedures and for the family to exercise an informed consent..

      The following challenges were discussed with committee members:

      1. development and testing of pediatric devices
      2. marketing of these devices in Canada.
      3. Health Canada's related policies (e.g. fee remissions), and whether Health Canada should be modelling approaches used in other jurisdictions.
2. Use of TAVI valves in the pulmonary position as opposed to the Melody valve, which has known problems (e.g. stent fracture) and is intended as a temporary device.

   Dr. Miro highlighted that any percutaneous valve and surgical bioprosthetic valves are temporary (Circulation. 2015; 131: 1960-70). He interpreted the Initial US experience as a learning curve that showed freedom from surgery was 87%, and freedom from re-intervention was 70% at 6 years. Most surgeries and re-interventions were related to fractures, and their incidence was greatly reduced by pre-stenting (HR=0.29).

   In a comparison of the Melody and Sapien valves, he indicated the initial and largest experiences were with the Melody valve, it offered easy deployment (BiB balloon), has a long landing zone, and is re-dilatable for re-valving. Drawbacks of the Melody include fracturing, endocarditis, necessity of a 22 Fr system (it is impossible to insert in patients less than 25 kg, unless hybrid approach), and the largest diameter of conduit/RVOT of 24mm.

   Possible advantages of the Sapien valve are: largest diameter conduit / RVOT of 29 mm; steerable delivery system of 16 Fr; fewer reported fractures; and lower incidences of endocarditis reported. Drawbacks of the Sapien valve include: single balloon deployment; shorter landing zone; less re-dilatable for re-valving.

   Dr. Miro noted that due to constant research investment in TAVI, it is likely that the Melody valve will be gradually replaced by devices with improved characteristics, and the current valves can only address less than half of patients that need valves. There is an on-going need for devices designed to treat patients with a dilated native RVOT.

3. Use of septal closure devices

   Dr. Miro indicated that the Amplatzer Atrial Septal Occluder (ASO) is the undisputed gold standard as it can close any size of ostium secundum defect, and almost any variation of anatomy in all sizes of patients. For this reason, Dr. Miro thought there was no significant demand for new devices. He indicated the one major drawback is erosion, which is somewhat predictable and occurs in 1/500 patients with a mortality rate of between 10-20%. Despite this, he indicated it is better than open heart surgery. The Gore GSO device does not have the same issue with erosion, but it cannot close medium or large ASDs. To be a credible contender device, it should treat the majority of patients, offer minimal or no risk of erosion, and be implantable in other types of ASD.

   Amplatzer devices can close almost any variant of Patent Ductus Arteriosus (PDA) anatomy in all sizes of patients. Currently available are the ADO I, ADO II, ADO II AS, ACP II, muscular ventricular septal defect (mVSD) devices for large PDAs and coil-based devices, Nit-Occlud and in small to medium PDAs. Therefore, there is not much need for new devices for PDA closure. Also, it's not clear what new advantage a device would offer and why you would need any PDA devices via the special access program.

   Although the Amplatzer muscular VSD devices can close most defects of this type, delivery systems are too large and stiff for smaller patients. Some special access justifications include the need for softer devices with smaller and steerable delivery systems.

   Dr. Miro indicated that there are no specific devices licensed in developed countries for membranous VSD closure. The incidence of AV block with mVSD1 is 5%, but the mVSD2 has shown excellent results in 38 patients between 2011 and 2014, but development was halted due to financial reasons. The device with characteristics adapted to the membranous septum, but with a lower incidence of AV blocks is the Lifetech CERA.

4. What evidence might be required for percutaneous interventions that simply 'buy some time' in patients who might need multiple open cardiac surgeries over their lifetime?

   Dr. Miro stressed that "buying some time" when you are a very sick before next surgery is not trivial. For some patients, buying time is not trivial, and avoiding an open heart surgery is not trivial, as the number of re-operations must be considered; The mortality rate increases with every re-operation and it is approximately 12% with the 5th surgery.

   Therefore the same evidence that is required for regular interventions; safety and efficacy in the short term (30 days) and the mid-term (1 year) is required. There are less concerns regarding long term evidence because these devices will need to be replaced, as they are not lifelong devices. They are being inserted to "buy some time".

Dr. Miro concluded his presentation.

10. Next Steps, Closing Remarks and Adjournment of Meeting

Dr. John Ducas, Committee Chair

The Chair thanked committee members and invited speakers for their participation and valuable input. Members will be canvassed to select a date for the next meeting.

Meeting adjourned.