This article presents a perspective casting doubt on the clinical and cost-effectiveness of trans-catheter aortic valve implantation (TAVI), and of European regulation of medical devices in general.1

The authors concede that severe aortic stenosis is the commonest cause of symptomatic valve disease in the elderly, that many of these patients are unsuitable for conventional aortic valve replacement, and that untreated most will be dead within 5 years. The Euroheart survey identified a large group of patients with significant aortic valve disease who were not offered valve intervention due to combinations of co-morbidities and age.2 The medically treated cohort of the Placement of AoRTic traNscathetER valve (PARTNER) trial demonstrated a 1-year mortality of 50.7% in patients who were deemed inoperable.3 Clearly, this is a patient group poorly served by conventional medical treatment; this goes some way to explain the enthusiasm to deliver an apparently effective treatment.

PARTNER is so far the only prospective randomised controlled study of TAVI versus optimal medical therapy including balloon aortic valvuloplasty (Cohort B) and aortic valve replacement (Cohort A). The authors cast doubt on the strength of evidence for TAVI presented by the PARTNER trial, but their presentation of the data is selective and lacks objectivity. In the 385 patient Cohort B study the authors are unimpressed by the 20% absolute reduction in mortality at 1 year with TAVI compared to optimal medical therapy. We believe that evidence for a therapy which saves 1 life for every 5 patients treated should not be disregarded because of minor differences in patient characteristics or the results of a smaller 90 patient follow-on study. In Cohort A the authors accurately describe the non-inferiority of TAVI compared to surgical AVR for the primary end-point of 12-month mortality, but with a higher rate of stroke and major vascular complications with TAVI, and a higher incidence of major bleeding and new onset atrial fibrillation in the surgical group. They fail to report that 30-day mortality according to actual treatment, conventionally considered procedure-related, was lower with TAVI, or that both symptoms and quality of life were significantly more favourable with TAVI at 30 days, only reaching equipoise at 6 months. This more rapid recovery post-procedure is of major importance to patients, particularly the elderly. It should also be noted that this was a trial of TAVI in cardiothoracic centres that had limited experience of the technique but outstanding expertise in surgical AVR.

The results of the PARTNER trial are supported by a number of major registry studies indicating outcomes with TAVI far better than predicted from estimates of surgical risk, and we believe provide a firm evidence base for TAVI in inoperable and high-risk surgical patients, as concluded by the FDA.

Furthermore we are not persuaded by the lack of evidence of an economic benefit. Three separate studies, including one by the authors themselves, have demonstrated evidence of cost-effectiveness in both inoperable and high-risk surgical patients.4-6

We respect the authors’ caution with respect to innovations and fully endorse their requests to strengthen the European system of licensing medical devices. We also agree that the rate of TAVI implants in certain European countries appears higher than the current evidence base justifies. However, we would caution against the extreme approach implemented by the FDA, which we believe denies patients access to safe and effective new treatments. It is inevitable that new surgical and interventional procedures for conditions that carry a high risk of death will have a limited evidence base at the time they are introduced. Inoperable patients not offered TAVI suffer a one year mortality of 50%; this is hardly analogous to the introduction of a new breast implant or hip prosthesis where effective treatment already exists. To some extent, therefore, practice will always move ahead of the evidence and early clinical experience will shape how the evidence is gathered. Only when there is enough clinical experience of the procedure to estimate the clinical benefit can the power calculations needed to design a randomized control study be performed.

In our experience of more than 200 TAVI procedures, patients with inoperable, symptomatic aortic stenosis have little difficulty understanding this.

Practice in the UK practice reflects guidance from the National Institute of Clinical Excellence (NICE) in that TAVI is offered only to patients deemed unsuitable for surgery, after thorough assessment by a heart team including a cardiac surgeon. The authors calculated that Belgian health authorities should be expected to pay the extra costs for TAVI in only 10% of patients considered for aortic valve intervention. In the UK, TAVI is commissioned at levels consistent with these recommendations at the rate of between 15-25 per million. All data is collected in the mandated UK TAVI registry, a valuable resource that is already shaping practice.

The title ‘TAVI: risky and costly’ is not applicable to UK practice and is irresponsible since unnecessary alarm may be felt by patients who are currently being assessed for a TAVI procedure. In the UK, TAVI has been introduced in a controlled and cautious way with full reporting of all outcomes to a national database. Randomised studies of TAVI against surgical aortic valve replacement are planned. Currently, however, the only indication for a TAVI procedure in the UK is severe symptomatic aortic stenosis in a patient considered to be too high risk for surgery by one or more cardiac surgeon(s). Our interpretation of the current evidence base would be: “TAVI in the UK – considerably less risky than untreated aortic stenosis.”

References.

1. Van Brabandt H, Neyt M, Hulstaert F. Transcatheter aortic valve implantation (TAVI): risky and costly. Bmj 2012;345:e4710.
2. Iung B, Baron G, Butchart EG, Delahaye F, Gohlke-Barwolf C, Levang OW, et al. A prospective survey of patients with valvular heart disease in Europe: The Euro Heart Survey on Valvular Heart Disease. European Heart Journal 2003;24(13):1231-43.
3. Leon MB, Smith CR, Mack M, Miller DC, Moses JW, Svensson LG, et al. Transcatheter aortic-valve implantation for aortic stenosis in patients who cannot undergo surgery. N Engl J Med 2010:363(17):1597-607.
4. Neyt M, Van Brabandt H, Devriese S, Van De Sande S. A cost-utility analysis of transcatheter aortic valve implantation in Belgium: focusing on a well-defined and identifiable population. BMJ Open 2012;2(3).
5. Watt M, Mealing S, Eaton J, Piazza N, Moat N, Brasseur P, et al. Cost-effectiveness of transcatheter aortic valve replacement in patients ineligible for conventional aortic valve replacement. Heart 2012;98(5):370-76.
6. Reynolds MR, Magnuson EA, Wang K, Lei Y, Vilain K, Walczak J, et al. Cost-Effectiveness of Transcatheter Aortic Valve Replacement Compared With Standard Care Among Inoperable Patients With Severe Aortic Stenosis / Clinical Perspective. Circulation 2012;125(9):1102-09.

Competing interests: All three authors participate in regular heart team meetings and implant the Medtronic Corevalve in suitable patients. Dr Malkin has implanted the Edwards Sapien valve in other centres. Dr Malkin has received travel grants from Medtronic (manufacturer of CoreValve aortic prosthesis). Dr McLenachan is the national clinical lead for specialist commissioning in the UK. Dr Blackman has received travel grants from Medtronic and receives remuneration for teaching and proctoring in other centres.

Since the first transcatheter aortic valve implantation (TAVI) was performed ten years ago, there has been a heightened interest in this evolving field amongst interventional cardiologists and cardiac surgeons.[1] Despite exponential growth in its utilization in Europe and United States, there remains a paucity of robust clinical evidence for this relatively novel technique. Van Brabandt and colleagues highlighted the lax and outdated approval processes for medical devices in Europe, exemplified by the comparison of quality assessment requirements for the Edwards Sapien valve and the Medtronics CoreValve to those of a domestic toaster.[2]

When considering interventional treatment of patients with severe aortic stenosis, the patient population should be considered in three subgroups according to their surgical risk for conventional aortic valve replacement (AVR): low surgical risk patients, high surgical risk patients, and inoperable patients, as demonstrated by Figure 1. A fourth subgroup of patients who are considered too unwell for any interventional treatment has also been represented. The available evidence for low surgical risk patients was examined in the STACCATO trial, which was prematurely terminated by a data safety monitoring board after unexpectedly poor results for patients in the TAVI arm.[3] The authors of the STACCATO trial modified their patient selection criteria mid-trial after initial results suggested superior AVR outcomes for lower surgical risk patients. For patients deemed inoperable by conventional AVR, the PARTNER Cohort B trial has demonstrated superior survival and symptomatic outcomes for patients who had TAVI compared to medical therapy, at a cost of an increased incidence of stroke.[4] However, the extent of benefits from TAVI may have been over-represented due to an incomplete analysis of data, as suggested by Van Brabandt.[2]

The most important question in the current clinical setting remains to be the comparative outcomes of patients who are deemed to have a high risk for AVR, but are nonetheless eligible surgical candidates. For these patients, it is important to recognize that the very definition of ‘high risk’ is heterogeneous and ill-defined amongst different institutions. Data from the PARTNER Cohort A trial suggests there were no significant differences in overall mortality, but a higher incidence of stroke/TIA after TAVI compared to AVR.[5] In addition, perioperative complications such as vascular injuries, permanent pacemaker insertion and paravalvular leaks have also been shown to be significantly more common after TAVI compared to AVR.[5, 6] Most importantly, there is no published comparative data on TAVI versus AVR with a follow-up of more than 5 years. From this point of view, we agree with Van Brabandt’s conclusion that current clinical practice has outpaced established evidence. Although TAVI has the potential to revolutionize the treatment of patients with severe aortic stenosis, its utilization should be restricted to inoperable patients until more robust long-term evidence is established.

REFERENCES
1. Cribier A. Development of transcatheter aortic valve implantation (TAVI): a 20-year odyssey. Arch Cardiovasc Dis. Mar 2012;105(3):146-152.
2. Van Brabandt H, Neyt M, Hulstaert F. Transcatheter aortic valve implantation (TAVI): risky and costly. BMJ (Clinical research ed.). 2012;345:e4710.
3. Nielsen HH, Klaaborg KE, Nissen H, Terp K, Mortensen PE, Kjeldsen BJ, Jakobsen CJ, Andersen HR, Egeblad H, Krusell LR, Thuesen L, Hjortdal VE. A prospective, randomised trial of transapical transcatheter aortic valve implantation vs. surgical aortic valve replacement in operable elderly patients with aortic stenosis: the STACCATO trial. EuroIntervention. May 14 2012.
4. Leon MB, Smith CR, Mack M, Miller DC, Moses JW, Svensson LG, Tuzcu EM, Webb JG, Fontana GP, Makkar RR, Brown DL, Block PC, Guyton RA, Pichard AD, Bavaria JE, Herrmann HC, Douglas PS, Petersen JL, Akin JJ, Anderson WN, Wang D, Pocock S. Transcatheter aortic-valve implantation for aortic stenosis in patients who cannot undergo surgery. N Engl J Med. Oct 21 2010;363(17):1597-1607.
5. Smith CR, Leon MB, Mack MJ, Miller DC, Moses JW, Svensson LG, Tuzcu EM, Webb JG, Fontana GP, Makkar RR, Williams M, Dewey T, Kapadia S, Babaliaros V, Thourani VH, Corso P, Pichard AD, Bavaria JE, Herrmann HC, Akin JJ, Anderson WN, Wang D, Pocock SJ. Transcatheter versus surgical aortic-valve replacement in high-risk patients. N Engl J Med. Jun 9 2011;364(23):2187-2198.
6. Wenaweser P, Pilgrim T, Kadner A, Huber C, Stortecky S, Buellesfeld L, Khattab AA, Meuli F, Roth N, Eberle B, Erdos G, Brinks H, Kalesan B, Meier B, Juni P, Carrel T, Windecker S. Clinical outcomes of patients with severe aortic stenosis at increased surgical risk according to treatment modality. J Am Coll Cardiol. Nov 15 2011;58(21):2151-2162.

Competing interests: None declared

Analysis of all randomised patients is required

We agree with van Brabandt and colleagues that full disclosure is required concerning patients with inoperable aortic stenosis who consented to be randomised to ‘trans-aortic valve insertion’ (TAVI) or ‘standard therapy’ as part of the PARTNER clinical trial (RCT).[1] This is essential so that other patients, clinicians and policymakers can be in a position to make informed decisions based on all the available evidence.

In the absence of such evidence we have combined the limited data from the US Food and Drug Administration (FDA) ‘briefing document’ (table 20) relating to the randomised 'continued access trial” (www.fda.gov) with the original published RCT data.[2,3] Our table shows the relative and absolute effect sizes based on this data. In order to calculate the relevant 95% confidence intervals around the benefits/harms of TAVI we used the actual numbers of events available for both trials. A caveat to this approach is that the published RCT reports outcomes at one year, whereas the “continued access trial” relates to outcomes over a shorter mean follow-up of 0.6 years.

The data currently in the public domain is insufficient to undertake the same analysis based Kaplan-Meier (KM) survival rates, but since the average follow-up for ‘all patients combined’ equates to 0.92 years we suspect that an alternative approach based on KM survival analysis would yield comparable results. For example based on the KM survival rates at 1 year in the “continued access trial” the relative risk for death equates to 1.59 with an NNTBenefit of 8, whilst the NNTHarm for stroke is 40; values which are similar to those shown in our table.

Analysis of the combined data in our table indicates that on average among every 100 patient receiving TAVI, rather than standard treatment (which included valvuloplasty among 84% of participants in the published RCT), 12 additional individuals will still be alive after one year who otherwise would have died on standard treatment. This benefit is at a cost of 4 additional individuals suffering a major stroke over the same time period. These results compare with 19 additional survivors and 4 people experiencing a major stroke after one year based on the published RCT data.[2] The combined data suggest that the ability of TAVI to reduce the risk of death is lower than that previously reported in the original RCT, but that the risk of major stroke is broadly similar.[2] Unfortunately it was not possible to assess the risk of ‘all strokes’ combined as the FDA data does not include ‘minor stroke’.[3] There is also a small inconsistency in the FDA data that requires clarification. The tables in the FDA report relate to 49 patients receiving ‘standard treatment’, but the related text refers to 50 patients randomised to ‘standard treatment’.[3]

van Brabandt and colleagues also raise concerns about differences between the two groups at baseline which may have led to an over-optimistic assessment of the survival benefits associated with TAVI.[1] Similar concerns were also raised when the original RCT was published and the trial investigators reported that “after adjustment for baseline risk imbalances, there were still marked differences in the mortality end point between the test and control therapies”, but did not provided details of the adjusted mortality hazard ratio and the related regression model.[4] Full disclosure of this multivariable model is now required. It would also be helpful if the same multivariable model could also be applied to the combined dataset comprising the published RCT and the “continued access trial” data.[2,3] The optimal scenario for patients, clinicians and policymakers would be for this further analysis to be published in a peer review journal according to the CONSORT criteria.

Michael A. Crilly
Senior Lecturer in Clinical Epidemiology
Aberdeen University Medical School, Polwarth Building, Aberdeen, Scotland. AB25 2ZD
mike.crilly@abdn.ac.uk

Christopher Littlejohn
Specialty Registrar in Public Health
NHS Grampian, Summerfield House, Aberdeen, Scotland, AB15 6RE.

CONFLICTS OF INTEREST
Both authors work for NHS Grampian and have advised the Health Board concerning the provision of TAVI. Both authors have also been involved in assessing whether individual patients with inoperable aortic stenosis should be referred for TAVI.

REFERENCES

[1] Van Brabandt H, Neyt M, Hulstaert F. Transcatheter aortic valve implantation (TAVI): risky and costly. BMJ. 2012;345:e4710. doi: 10.1136/bmj.e4710.

[2] Leon MB, Smith CR, Mack M, Miller DC, Moses JW, Svensson LG, et al; PARTNER Trial Investigators. Transcatheter aortic-valve implantation for aortic stenosis in patients who cannot undergo surgery. N Engl J Med. 2010;363:1597-607.

[3] Food and Drug Administration (FDA). The Edwards SAPIEN® THV Transcatheter Heart Valve System for Patients With Severe Aortic Stenosis Who Are Not Candidates for Conventional Open-Heart Aortic Valve Replacement Surgery. Briefing Document for the Circulatory Systems Device Panel Advisory Committee. Other Experience in Inoperbale Patients. Randomized, Continued-Access Patients. FDA (20 July) 2011: 95-99.

[4] Leon MB, Smith CR, Tuzcu EM. Transcatheter Aortic-Valve Implantation for Aortic Stenosis.
N Engl J Med 2011; 364:179-181.

Competing interests: Both authors work for NHS Grampian and have advised the Health Board concerning the provision of TAVI. Both authors have also been involved in assessing whether individual patients with inoperable aortic stenosis should be referred for TAVI.

This nice study adequately points to 2 fundamental problems in surgery.

Medical devices are introduced and used without adequate clinical evaluation. Generally only safety has been demonstrated, whereas the clinical use is the responsibility of the surgeon. In gynaecology - we will restrict ourselves to the area we are familiar with - this can be illustrated by the use of meshes in pelvic floor surgery. The use of meshes during laparoscopic reconstruction was initiated in Europe, pioneered by the French in the early nineties. The introduction in the US followed many years later, mainly because of concerns of the surgeon’s responsibility. We witnessed the widespread introduction of vaginal meshes for urinary incontinence and for pelvic floor descent. Clinically we have become aware of some safety aspects as large pore sizes and low weight meshes while complications as mesh erosion have been documented. This has been discussed widely at meetings and in the literature. Today almost 20 years later, concern about complications is growing.

Barbed sutures were introduced 5 years ago as an innovative suture material facilitating running sutures without the necessity of laparoscopic knot tying. Over the last years several case reports of associated small bowel occlusion were presented at meetings some of which were published1-3.

These examples highlight both the absence of large scale clinical trials before the introduction of new medical devices, and the post marketing medical surveillance. The latter however is slow and poorly organized, although discussed repetitively at workshops and meetings. It moreover is not very effective since we all refrain from too strong statements by lack of “solid” evidence and medico legal implications.

The major underlying problem however, is the overall absence of quality control in surgery. In contrast with the large RCTs and post marketing surveillance of drug treatment, the quality control of individual surgical interventions is close to inexistent. Video registration of entire surgical interventions, if introduced for this purpose, will probably enhance quality while reducing costs. The obvious advantages for society and health care cost however are opposed because of medico legal and other concerns. In surgery indeed the skill of the surgeon and aspects as circumstances, equipment and fatigue, besides technique play a major role in the quality of the individual interventions and in the eventual complications. Without mandatory video registration4 of entire interventions, the problems will persist.

Reference List

(1) Donnellan NM, Mansuria SM. Small bowel obstruction resulting from laparoscopic vaginal cuff closure with a barbed suture. J Minim Invasive Gynecol 2011; 18(4):528-530.
(2) Patri P, Beran C, Stjepanovic J, Sandberg S, Tuchmann A, Christian H. V-Loc, a new wound closure device for peritoneal closure--is it safe? A comparative study of different peritoneal closure systems. Surg Innov 2011; 18(2):145-149.
(3) Thubert T, Pourcher G, Deffieux X. Small bowel volvulus following peritoneal closure using absorbable knotless device during laparoscopic sacral colpopexy. Int Urogynecol J 2011; 22(6):761-763.
(4) Koninckx PR. Videoregistration of surgery should be used as a quality control. J Minim Invasive Gynecol 2008; 15(2):248-253.

Competing interests: CEO EndoSAT NV