'Substantially Equivalent'; Good Enough for Lower-Limb Therapy Patients: Contributions to the Meaningful Data Collection Discussion

This letter is a response that elaborates on the peripheral artery disease (PAD) data collection strategy proposed by Dr. Lawrence Garcia in his blog post, “Collecting More Data for Vascular Patients.”¹

The root of the problem with PAD data collection lies in the scope of industry-funded PAD device trials, which are commonly designed for minimum United States Food and Drug Administration (FDA) 510(k) requirements — a relatively less complex and expensive process than post-market approval (PMA). This requires manufacturers to prove that a device is “substantially equivalent” in indications for use, technological characteristics, and safety and performance results compared to a previously marketed (predicate) device. Complicating matters, investigator initiated post-market studies funded by industry are predominantly concerned with new indications with similar limitations.

Industry funded trials for regulatory clearance or new indications generally examine relatively small, niche patient populations. There is a notable mismatch between industry-funded research hypotheses and clinically/economically relevant solutions for integrating PAD devices into the clinical care pathway. For instance, 510(k) clearances for TurboHawk Peripheral Plaque Excision (Covidien) and Stealth 360° Orbital PAD Systems (Cardiovascular Systems, Inc.) were based solely on bench and biocompatibility testing. While the 2004 clearance of the CLiRpath Excimer laser catheter (Spectranetics) required clinical evidence, only 47 patients were examined and adverse events were assessed arbitrarily against a previous registry study. The 2005 Spectranetics 510(k) clearance required only bench and biocompatibility testing. The 2014 clearance of the Phoenix Atherectomy System (AtheroMed; predicate devices included ev3 SilverHawk Peripheral Plaque Excision System and Pathway Medical Jetstream System) required clinical study of major adverse events over only the first 30 days in 123 patients, and it received clearance even without any long-term follow-up.

Pilot trials produce relatively scarce evidence that is grossly inadequate to realistically predict clinical outcomes. Industry-funded trials supporting PMA for the IN.PACT Admiral drug-eluting balloon catheter (Medtronic) consisted of 29 studies including more than 4,600 patients at 230 sites.² Approval for the Zilver PTX Drug-Eluting Peripheral Stent (Cook Medical), the first drug-eluting stent (paclitaxel) for PAD,³ required both nonclinical data and multiple clinical trial programs of 479 patients over 12 months, 787 patients over 12 months, and an FDA-mandated additional 5-year, 900 patient post-approval study.³ Thus, the onus remains on practitioners and researchers to estimate long-term device performance and outcomes for 510(k) devices, with limited data-driven quality comparison studies and few comparative effectiveness research (CER) studies for support. 

Fundamentally, I embrace Dr. Garcia’s notion that a large body of rigorously collected data from lower-limb therapy interventions is urgently needed. The concept of a “steering committee” serving as the core lab, ideally unbiased by financial considerations of the study, is noble, but may not be practicably executed. Rather than eliminating or limiting industry involvement, which innately placed limits on funding and resources, an alternative strategy is working through third-party entities receiving industry and academic input (e.g., National Heart, Lung, and Blood Institute, The American Heart Association, and American College of Cardiology Foundation). These entities can equitably distribute funds while facilitating transparency, collaboration between competitors, and pragmatic trial designs that focus on the real issue of identifying populations that will benefit from each interventional device (rather than head-to-head comparisons of device inferiority).⁴ 

While a large randomized prospective trial, as suggested in the double-randomization approach suggested by Dr. Garcia, is desirable, it may be more costly, heavily impacted by device lifecycle (phasing out of devices and entry of new devices to market), and likely to produce results that inadequately represent the general population than a collaborative and ongoing structured clinical registry effort for PAD.

Contemporary software solutions can allow for unprecedented effectiveness in multi-institution data pooling, 4 evidenced by cardiovascular registries, including the CONCOR registry for congenital heart disease, Acute Coronary Syndrome (ACS)/National Cardiovascular Disease Database (NCVD), and eight active National Cardiovascular Data Registry (NCDR; www.ncdr.com) registries.⁵ Notably, NCVD recently expanded the existing CARE Registry to the Peripheral Vascular Intervention (PVI) Registry in 2014. Contemporary sophistication and utility of collaborative clinical registries also allow for custom queries, patient data comparisons (e.g., procedure number and type), site information (i.e. teaching vs nonteaching, location) on a much larger scale than feasible for randomized studies. Furthermore, as devices are cleared for use, registry programs are already in place for post-market assessments.

Because both diagnostic and interventional technologies for PAD treatment and presentation are rapidly advancing, inconsistencies in practitioner training and skill maintenance also pose challenges and cause bias in clinical trials, complicated by specialized training often provided by industry sponsors. The “data set driven by the best and brightest of operators” suggested by Dr. Garcia in actuality may not accurately represent how PAD devices are performing under routine clinical care situations. Plus, qualification of the “best and brightest” without extensive and costly training and assessment programs may be prohibitively challenging for a variety of modern PAD interventional devices, resulting in bias due to preference of the practitioners involved.

Similarly, minimizing inclusion and exclusion criteria to achieve a larger patient population can open large-scale studies to bias caused by site-based variations in the effectiveness of PAD screening and prevention programs. While a large dataset is desirable to explore correlations between rare events and specific characteristics of patient populations, prospective collection and control for potentially confounding variables is necessary to ensure accurate and minimally biased data representation.

As PAD interventional technologies mature, both prospective randomized trials and extensive large-scale collection of multi-institutional information must be utilized to collect the information on outcomes and performance necessary to advance patient care. Future improvements in the PAD care pathway will require leveraging the contribution of multi-institutional technology systems to create viable means of collecting information about large numbers of PAD patients in a sustainable and ongoing manner. Such data will be required to move beyond token PAD device pilot trial results and provide objective evidence of which devices are more than merely “substantially equivalent” in real clinical populations. 

Editor’s Note: Disclosure: The author has completed and returned the ICMJE Form for Disclosure of Potential Conflicts of Interest. The author reports no conflicts of intersted related to the content herein. 

Manuscript submitted February 17, 2014; final version accepted February 24, 2014. 

Address for correspondence: Raghotham R. Patlola, MD, FACC, Cardiovascular Institute of the South, Division of Cardiology, 2730 Ambassador Caffery Parkway, Lafayette, LA 70506, United States. Email: rpatlola@aol.com.

References

1. Garcia LA. Collecting more data for vascular patients. Vasc Dis Manag. 2013;April 15. https://www.vasculardiseasemanagement.com/blog/collecting-data-atherectomy-patients 
2. Medtronic, Bard in close race to first FDA-approved drug-eluting balloon. Medical Devices Today. 2013, June. https://www.medicaldevicestoday.com/2013/06/medtronic-bard-in-close-race-to-first-fda-approved-drug-eluting-balloon.html
3. Suvarna V. Investigator initiated trials (IITs). Perspect Clin Re
4. s. 2013;3(4):119–121.
5. United States Food and Drug Administration. FDA approves first drug-eluting stent to treat peripheral arterial disease. November 15, 2012. https://www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/ucm327068.htm.
6. Williams WG, McCrindle BW. Practical experience with databases for congenital heart disease: a registry versus an academic database. Semin Thorac Cardiovasc Surg Pediatr Card Surg Annu. 2002;5:132-142.