Introduction of New Medical Devices: Lessons Learned From Experience With Endovascular Aneurysm Sealing

Editorials published in the Journal of Endovascular Therapy reflect the opinions of the author(s) and do not necessarily represent the views of the Journal, the International Society of Endovascular Specialists, or SAGE Publications Inc.

Innovation is the cornerstone of surgical practice. Consider the aortic stent-graft. After years of experimental and clinical research during the 1980s by Volodos and colleagues ¹ in Ukraine, endovascular aneurysm repair (EVAR) was introduced to the West in the early 1990s. ² It has since transformed the management of aortic aneurysm. The impressive 30-day mortality and short-term outcomes achieved with EVAR enabled the treatment of patients who would otherwise have been considered unfit for intervention. ³ Indeed, the short-term gains observed with elective infrarenal EVAR have been echoed in the IMPROVE trial, ⁴ which also concluded that EVAR was efficacious and cost effective for ruptured infrarenal abdominal aortic aneurysm (AAA).

With such significant improvement in short-term outcomes, medical device companies quickly began competing to bring new, innovative endografts to market. Concerns about the long-term durability of EVAR are not new,^5,6 but proponents have argued that historical data with old technology does not accurately reflect current practice. ⁷ The early adoption of new aortic stent-grafts has increased significantly; as of 2010, EVAR accounted for 74% of AAA repairs in the United States. ⁸

Fifteen-year follow-up data of the EVAR 1 trial has made for uncomfortable reading. Beyond 8 years there was a significant survival advantage for open AAA repair. ⁹ Furthermore, EVAR was associated with more reinterventions and a high burden of surveillance. These data have recently been corroborated by a large series from the United Kingdom of over 38,000 unruptured AAA repairs collected by the UK Hospital Episode Statistic. ¹⁰ EVAR was associated with higher 10-year mortality and higher risk of rupture over 9 years than open surgical repair (3.4% vs 0.9%).

Endoleak remains a major factor in reducing the long-term durability of EVAR. Types I and III endoleaks require prompt intervention, and type II endoleaks should be treated in the presence of significant sac enlargement to reduce the risk of secondary sac rupture. ¹¹ To address this complication, the revolutionary technique of endovascular aneurysm sealing (EVAS) was developed to obliterate the aneurysm sac. ¹² The Nellix EVAS system (Endologix Inc, Irvine, CA, USA) consists of balloon-expandable stent-grafts surrounded by endobags containing a polymer that was designed to fill the aneurysm sac and eliminate space for endoleak formation; there is no proximal fixation. The pivotal trial ¹³ investigating the safety and efficacy of Nellix EVAS reported encouraging 30-day and 1-year outcomes, which led to marketing approval in the United States. These trial outcomes were confirmed in larger real-world multicenter cohorts.^14,15 EVAS was reported to be quick to deploy, with reduced radiation exposure to patient and operating team.^16,17

Unfortunately, longer-term follow-up demonstrated significant failure rates beyond 2 years after EVAS implantation. Higher incidences of type Ia endoleak, graft migration, and secondary sac rupture were reported in the EVAS FORWARD investigational device exemption (IDE) trial, ¹⁸ which resulted in refinement of the instructions for use (IFU). Significant midterm failures were reported by Harrison et al ¹⁹ at a large-volume UK tertiary center. In this cohort, device failures were seen in 26% of patients, rates alarmingly higher than for EVAR. Further data from United Kingdom and Europe supported these findings.^20,21 In light of this and in consultation with the Medicines and Healthcare Products Regulation Agency, Nellix was withdrawn from the UK market.

Interestingly, in vivo mathematical geometric modeling suggested that the forces created by the mass of the Nellix endobags in the aneurysm sac were significantly higher than in EVAR or open repair. ²² In the absence of proximal or distal fixation, the endobags could dislodge during routine activities of daily living, which was predictable from mathematical modeling. With further emerging data regarding midterm EVAS failure, ²³ it is important to understand the regulatory processes involved in bringing a new medical device to market.

In the United States, for a clinical trial where there is potential for significant risk to the patient, protocols need to be approved by the Food and Drug Administration and evaluated under an IDE. After the data have been analyzed, the manufacturer can apply for market approval, typically based on 12-month outcomes; after this, the device has the potential to be widely adopted. This system effectively bases device safety decisions on 30-day outcomes and device efficacy at 1 year. It is clearly not designed to monitor long-term durability. In Europe, the Conformité Européenne mark can be placed on a product when it has met the requirements in the Medical Device Directive, at which point it can be freely marketed in the EU. This is usually based on a clinical trial but data from in vitro and in vivo work can also be used in support. Again, short-term safety and efficacy are the primary focus.

EVAS is certainly not the first novel technology to be withdrawn from the market as a result of high failure rates following its commercial introduction. Vascular surgeons have seen a number of aortic stent-grafts come to the market to be later withdrawn. Indeed, device durability is important across the spectrum of medical devices; many will recall the reports of problems with metal-on-metal hip prostheses and breast implants over the last few years. The senior author had significant personal experience of the silicone-based Angelchik device designed to reduce gastroesophageal reflux in patients with hiatus hernia. The device was quickly adopted across several institutions and was marketed as a minimally invasive technique that would revolutionize the management of gastroesophageal reflux in the 1990s. As a surgical trainee, the senior author assisted in the adjustment and removal of these prostheses. However, it soon became clear that failure rates through migration, erosion, fistula formation, and dysphagia were much higher than expected. ²⁴ Despite these early warnings the device continued to be used. After several case series were published showing poor outcomes, Angelchik was withdrawn from the market.

There are many parallels between the Angelchik and Nellix devices, with early promise of potentially game-changing technology undermined by durability issues and poor longer-term outcomes. Over the decades since the introduction of the Angelchik, there has been little change in regulations governing the introduction of new medical devices. There is an important balance to be struck between interests of an individual patient and those of many future patients who may benefit from innovation and product development. It may be argued currently, however, that the pendulum has swung too far in the direction of innovation at the expense of patient safety.

The current focus on 30-day and 1-year outcomes to support safety and efficacy is clearly not sufficient or fit for the purpose. The real question now is what can we do to prevent these issues in the future? National and international registries that focus on recording long-term outcome data could be the answer. Periprocedural mortality for elective EVAR is so low that it is no longer a valid outcome measure. We should be focusing our efforts on recording reintervention rates after EVAR, which occur in approximately a fifth of patients. ²⁵ Registries can hold data on devices inserted and identify early those that are not performing well. The orthopedic surgeons in the United Kingdom have done this well through the National Joint Registry, which was able to identify early the high revision rates associated with metal-on-metal hip implants. ²⁶ This is the type of registry model we should be developing in vascular surgery, with a focus on accurate recording of device information at the time of implantation, including patient anatomical variables and a subsequent endpoint of reintervention rather than 30-day or in-hospital mortality. If this system had been in place, then it is likely that problems with EVAS would have been identified earlier to allow enhanced surveillance and reintervention at an earlier stage and to reduce the numbers of patients treated with this device.

Clearly there is a need to change our current practice, and it appears there is appetite for this from key stakeholders. Perhaps it was the race to be at the forefront of technology that led to increased application of EVAR, often outside the IFU. Whatever the impetus, the UK’s National Institute of Clinical Excellence (NICE) has drafted guidelines that strongly favor open AAA repair over EVAR. ²⁷

For a device to be approved for use, its safety and effectiveness need to be demonstrated. New technology has helped to improve short-term outcomes for AAA patients but can result in failure patterns in the longer term that were not appreciated at the inception of EVAR. While we must be careful not to stifle innovation we must ensure patient safety is not compromised in pursuit of the next paradigm shift in technology. Through collaboration and collection of device-specific information and long-term data in national and international registries we can ensure the safe introduction of new technology in a standardized manner so that ultimately patients are the ones who gain the greatest benefit from the introduction of new techniques.