Evoque and Patisiran: A Tale of Two Regulatory Journeys

Drugs and devices travel different regulatory paths to approval: can evidentiary rigor and regulatory flexibility coexist?

Evoque and Patisiran: A Tale of Two Regulatory Journeys

Two decisions by the US Food and Drug Administration in recent months have got me thinking about how to balance the need for evidentiary rigor with the need for regulatory flexibility enabling timely access to life-improving or life-extending treatments.

Last month, the FDA approved the Evoque transcatheter tricuspid valve replacement system (Edwards Lifesciences) for the improvement of health status in patients with symptomatic severe tricuspid regurgitation (TR) despite optimal medical therapy (OMT) for whom tricuspid valve replacement (TVR) is deemed appropriate by a heart team.

Four months earlier, the FDA had declined to approve patisiran (Onpattro; Alnylam Pharmaceuticals), an RNAi therapeutic, for the treatment of patients with transthyretin amyloid cardiomyopathy (ATTR-CM) despite the trial meeting its primary endpoint of difference in 6-minute walk distance (6MWD) from baseline to 12 months in the APOLLO-B pivotal trial

Evoque’s approval was supported by preliminary data from the pivotal trial TRISCEND IIresults for the first 150 out of 392 patients were presented at TCT 2023. In the early phase data reported during the meeting, the trial met its primary safety endpoint—a composite of major adverse events at 30 days using a performance goal of 70%—and its two co-primary effectiveness endpoints—a TR grade reduction to moderate or less at 6 months (98.8% vs 21.6%, P < 0.001) and a hierarchical composite effectiveness endpoint of KCCQ, NYHA, and 6MWD improvement from baseline to 6 months (P < 0.001). A win-ratio analysis was also supportive. At the time of the approval, the sponsor noted that 318 patients had completed 1-year follow-up and that there are “favorable trends” in support of the primary endpoint. Of note, the device was approved without first being put through an FDA advisory committee (AdComm) meeting.

By contrast, patisiran had been subject to FDA AdComm review, with the panel voting 9-3 that the benefits of the drug outweigh its risks. APOLLO-B, supporting that vote, was a double-blind, placebo-controlled trial, with more than double (n = 359) the number of patients available for FDA review, longer follow-up (12 months), and thanks to the blinded design, an unbiased assessment of 6MWD (the primary endpoint) and health status. In its Complete Response Letter to the sponsor, the FDA noted that the “clinical meaningfulness” of patisiran’s treatment effects for ATTR-CM could not confirmed. Based on the FDA’s decision, the sponsor is no longer pursuing an expanded indication for patisiran in ATTR-CM in the US.

The different paths and outcomes of these two therapies highlights the marked contrast between device and drug approval processes. Devices, in general, do not require the same level of evidence as drugs. Conducting clinical studies for implantable devices is often more complicated and expensive than for drugs and can be ethically problematic with regards to placebo (sham) control. Unlike the “substantial evidence of safety and effectiveness” standard for approval of drugs (and biologics) which requires a statistically persuasive (P < 0.001) and clinically meaningful benefit in adequate and well-controlled trial(s), the standard for marketing a medical device is “reasonable assurance of safety and effectiveness” (RASE).

Only high-risk Class III devices that are lifesaving or life-supporting, such as Evoque, are required to submit clinical data to demonstrate RASE via the premarket approval (PMA) pathway. PMA is based on sufficient valid scientific evidence, which unlike the drug approval standard, is broadly defined ranging from well-controlled investigations to case reports or significant human experience with a marketed device.

Alleviating a Burden

In recent years, the US Congress has directed the FDA to take a “least burdensome” approach, hoping to speed timely access to technologies that extend or improve life. That’s meant shifting the statutory burden of proof from exclusively preapproval to a lifecycle approach including postapproval evaluations. In addition, expedited approval pathways such as the Breakthrough Devices Program for life-threatening or irreversibly debilitating conditions for which no approved alternatives exist permit for more efficient and flexible trial designs. Here, any shortfall or uncertainty surrounding premarket evidence is intended to be bolstered by mandatory adequate postmarket studies and scrutiny.

The problem? The FDA-mandated postapproval studies doesn’t exactly have a stellar track record, with many such studies being delayed, scaled back, or never completed. And even though the FDA has increased Congress-legislated authority to revoke approval of drugs and devices that fail to confirm clinical benefit in postapproval studies, it rarely does so.    

It was the Breakthrough Devices Program that enabled Evoque’s pivotal trial to employ a phased primary analysis plan so that a health status improvement indication could be approved on the basis of the 6-month results of those first 150 patients. A second future PMA supplement could subsequently request the indication be expanded to include morbidity/mortality improvement based on the final results of all patients enrolled at 1 year.

February’s approval decision was thus based primarily on the former. The device met the primary safety endpoint, albeit against a lax performance goal of 70%. Whether the seemingly impressive reduction in TR grade is clinically meaningful benefit and capable of altering the natural history of the disease is not clear because the trial was unblinded: health status improvement was driven by outcomes that are susceptible to bias/placebo effect and more-objective measures, such as impact on NT-proBNP biomarker or intensity of medical therapy, were not reported. Open-label clinical trials, where patients and investigators are aware of assigned therapy, are rarely adequate to support labeling claims based on patient reported outcomes.

A Decade of Device Decisions

The approval of cardiac devices over the last decade illustrates the apparent laxity and capriciousness of medical device regulation, including many which were approved despite either equivocal or negative findings in their pivotal trials.

In October 2013, the MitraClip (Abbott) was approved for use in patients with primary degenerative mitral regurgitation (MR) at prohibitive risk for mitral valve surgery based on two high-risk registries (this group had been excluded from the pivotal randomized EVEREST II trial where the device was shown to be inferior to surgery). The indication was extended to secondary functional MR in 2019 despite contradictory results from industry-funded COAPT (positive) and government-funded MITRA-FR (neutral).

In May 2014, the CardioMEMs HF System was approved for NYHA class III heart failure (HF) based on the CHAMPION trial, having been rejected by an earlier panel in December 2011. The FDA disagreed with the panel’s vote against a reasonable assurance that the device was effective. The indication was extended for NYHA class II heart failure in 2022 based on a positive subgroup of GUIDE-HF, which had failed to meet its primary endpoint (HF hospitalizations, urgent HF visits, and mortality).

The Watchman left atrial appendage occlude (Boston Scientific) had previously been turned away by the FDA based on the mixed results of the PROTECT AF trial, only to be approved in March 2015 based on the PREVAIL trial (wherein the two effectiveness endpoints were not met and the safety endpoint was barely met). The agency’s signoff was for an indication not evaluated in either trial—for patients with a rationale for a nonpharmacologic alternative to warfarin.

Approved in March 2015 without an AdComm panel, the Impella 2.5 system (Abiomed) was indicated for improving hemodynamic support during high-risk percutaneous coronary intervention without increasing major adverse events, including mortality, compared to intra-aortic balloon pumps (IABPs) despite the pivotal PROTECT II trial being prematurely terminated for futility and safety concerns. Of note, IABPs are no longer recommended or have been downgraded by guidelines.

In July 2016, the Absorb bioresorbable scaffold (Abbott) was approved for improving coronary luminal diameter in patients with ischemic heart disease based on noninferiority at 1 year in ABSORB III, but outcomes at longer follow-up uniformly went in the wrong direction. The device was recalled from the market less than a year later due to increased risk of CV events.

The prespecified intention-to-treat analysis of the RESPECT trial found no statistically significant reduction in the primary endpoint of ischemic stroke among patients with cryptogenic stroke treated with the Amplatzer PFO occluder (Abbott), yet in October 2016 the FDA deemed the 50% stroke reduction to be clinically meaningful—leading to its approval based on a total of 25 events!

Back in March 2016, the FDA’s advisors voted 12-0 against recommending approval of the AngelMed Guardian (Avertix) implantable cardiac detection monitor and ACS warning system on the basis of efficacy and risk-benefit balance; two years later an extended follow-up analysis outside of the prespecified randomized period of the pivotal ALERTS trial led to FDA approval.

Both the Paradise (Recor Medical) and the Symplicity Spyral (Medtronic) renal denervation systems were approved for resistant hypertension a little over a week apart, in November 2023, despite the FDA’s AdComm’s endorsement of the former, but not the latter, which had missed its primary endpoint in the SPYRAL HTN-ON MED trial (although not in the OFF-MED trial). The magnitude of systolic BP reduction with these devices is a modest 4 to 6 mm Hg, which is about half the reduction seen with a single dose of a commonly used antihypertensive, aldactone 25 mg.

To summarize, two of these devices were approved without AdComm panel review. Of the eight remaining, the FDA advisors voted against approval in one instance, were decidedly mixed in four other meetings, and were strongly positive in three cases. Ironically, the only device to receive a unanimously positive vote (9-0 with 1 abstention) was the Absorb BVS—and we know how that ended. Of note, all 10 devices were approved with the caveat that postapproval studies were required; some of these are currently ongoing.

Setting the Bar

The months to come may add more complexity. At an AdComm meeting mid-February, less than 2 weeks after Evoque’s approval, the panel voted 13-1 in favor of the TriClip device (Abbott) for symptomatic severe TR, largely on the basis of the TRILUMINATE trial. An FDA decision is currently pending, but it seems likely that US physicians will have access to two new transcatheter devices intended for the same indication, both winning on primary safety and effectiveness endpoints. Both approvals were driven primarily by patient-related outcomes in an unblinded trial, raising questions about bias/placebo effect. And yes, only one device required an AdComm panel, while the other was approved without panel input!

All of these examples raise the question whether the current bar for FDA approval of new cardiac devices is too low, particularly when you consider the regulatory climate for medications. My worry is that the FDA’s new attempts at regulatory flexibility, at best, leads to inconsistent decision making and, at worst, erodes clinician and patient confidence in the FDA. Promoting innovation and protecting public health is best served by exercising sufficient caution before allowing these devices to be marketed—and continuing to follow them, rigorously, once in the hands of physicians and in the bodies of


Off Script is a first-person blog written by leading voices in the field of cardiology. It does not reflect the editorial position of TCTMD.

Sanjay Kaul, MD, is a cardiologist and professor of medicine at the Cedars-Sinai Medical Center in Los Angeles. He has…

Read Full Bio
  • Kaul reports consulting on most of the device examples (and the drug example) listed in this blog.