Confounders and Discrepancies: We Need More Rigorous Trials

Can you describe your reaction to the recent Katsanos et al meta-analysis¹ and FDA’s response thus far?
Any mortality concern regarding our patients is very important and when a study receives as much press as the Katsanos meta-analysis, our patients also express concern. Our primary goal as physicians is to do no harm. Currently we are all reviewing the data and the meta-analysis, and analyzing the conclusions of this study. I am concerned at the mortality discrepancy between the meta-analysis and what was reported individually in each of the randomized, controlled IDE trials: IN.PACT SFA (Medtronic), Lutonix LEVANT 2 (BD), and ILLUMENATE with the Stellarex DCB (Philips). These three randomized control trials did not show any difference in mortality. Another major concern regarding the meta-analysis is the way it was conducted. Some obvious concerns are that the meta-analysis initially included 28 randomized, controlled trials, but the primary conclusions were only based on the 3 randomized, controlled trials that had 5-year data. Other assumptions made in the meta-analysis may not be accurate, such as the dose relationship to mortality and the calculation of the dose delivered by drug-eluting stents. Also, this meta-analysis was not a patient-level meta-analysis and did not have any adjudication of events (mortality).

What raised the level of concern were the two FDA warning letters: the first letter² was in January 2019 (soon after the publication of the meta-analysis), where the FDA recommended the following:

• Continue surveillance of patients who have been treated with paclitaxel-coated balloons and paclitaxel-eluting stents per the current standard of care.
• In clinical decision-making, discuss the risks and benefits of all available treatment options for peripheral arterial disease (PAD) with your patients.
• Report any adverse events or suspected adverse events experienced with the use of paclitaxel-coated balloons and paclitaxel-eluting stents. Voluntary reports can be submitted through MedWatch, the FDA Safety Information and Adverse Event Reporting program. Device manufacturers and user facilities must comply with the applicable Medical Device Reporting (MDR) regulations. Health care personnel employed by facilities that are subject to FDA's user facility reporting requirements should follow the reporting procedures established by their facilities. Prompt reporting of adverse events can help the FDA identify and better understand the risks associated with medical devices.

An update to that letter³ came out on March 15, 2019, after the FDA did their own analysis, and was a much stronger letter that indicated definite concern. It noted that in total, among the 975 subjects in these three trials, there was an approximately 50% increased risk of mortality in subjects treated with paclitaxel-coated devices versus those treated with control devices (20.1% versus 13.4% crude risk of death at 5 years). In addition to the above recommendations, they added that clinicians should discuss the risks and benefits of all available PAD treatment options with their patients. The FDA also said that for most patients, alternative treatment options to paclitaxel-coated balloons and paclitaxel-eluting stents should generally be used until additional analysis of the safety signal has been performed.

This update of the FDA statement has created a wave of concern among physicians who are treating peripheral vascular disease. Some have responded by having open discussions with their patients regarding the study and others have stopped using paclitaxel-coated balloons until the FDA offers further guidance after the June meeting.

How much of this comes down to a debate about the quality of the data and the methodology of analysis?

All the randomized control investigational device exemption (IDE) studies were conducted with a data safety monitoring board and an independent clinical events committee. These studies were conducted very well, so there is no concern over the quality of the data or methodology of analysis. The trials were powered to look at patency and safety as the primary endpoint, and mortality as a secondary endpoint. Patency was the primary endpoint, because in peripheral arterial disease, the Achilles’ heel of the available non-drug-coated therapies is a very high reintervention rate, due to repeat restenosis in the treated lesions. The goal of the paclitaxel-coated balloons is to improve the primary patency and reduce the rate of repeat reintervention. Increased mortality with paclitaxel has not manifested elsewhere in its therapeutic use in medicine and it is used widely in breast cancer, where it has a survival benefit. Paclitaxel has been clinically approved for use in coronary artery disease where there was no mortality concern attributed to paclitaxel, so peripheral arterial disease trials were not powered for mortality as a primary endpoint. These trials were conducted with excellent rigor. In all trials, it is common to have patients lost to follow-up. In a large, randomized, controlled trial, it is not uncommon for patients to drop out, move, or not be compliant with follow-up out to 5 years. All these randomized control IDE paclitaxel trials were designed with a 2:1 randomization, meaning for every 2 patients receiving drug-coated balloons, 1 patient receives a percutaneous transluminal angioplasty (PTA) balloon. A 2:1 randomization was done in order to expedite the time to trial completion. This 2:1 randomization and the availabity of commercially available paclitaxel products make the conclusion of the meta-analysis suspect. At the time of the majority of the trials, with the exception of the Zilver PTX trial, paclitaxel was commercially available in the market. Remember that most of these patients have systemic atherosclerosis, which means that both of the legs potentially will be affected by the disease process, causing symptoms. It is likely that a patient was randomized into a paclitaxel-coated device trial and had one limb with symptomatic disease randomized to the balloon angioplasty arm. After a period of time (which varies, depending on the study protocol), the patient was eligible to undergo another  procedure on their other affected, symptomatic leg. That patient, more than likely, would have received treatment with an FDA-approved paclitaxel-coated balloon that was available in the market at that time. Thus, we are unlikely to have any paclitaxel-naïve patients, other than in the Zilver PTX study. That is a real factor, because if the conclusion of this meta-analysis is that paclitaxel causes mortality and if some of the patients in the balloon angioplasty arm received paclitaxel therapy to the contralateral limb, that conclusion may not be true, because we truly may not have any paclitaxel-naive patients. This circumstance creates a major problem with the mortality conclusions associated with paclitaxel in this meta-analysis. Therefore, we cannot make a broad assumption about paclitaxel’s link to mortality based on the meta-analysis, because in reality, the number of paclitaxel-naive patients may have been small at the time of follow-up, especially at 5 years.

Another important element is that dose calculations for paclitaxel in drug-eluting stents were incorrect. In a recent editorial⁴, Dr. Andrew Holden stated that Katsanos et al “incorrectly assumed that drug-eluting stents (DES) liberate paclitaxel over the entire lumen surface of the vessel. Holden et al note that only 14% of the abluminal surface of the Zilver PTX DES (Cook Medical) is coated with the drug, resulting in far less paclitaxel exposure than postulated by Katsanos in his meta-analysis.”⁵ Based on this incorrect assumption of dose, the Katsanos meta-analysis indicated a relationship between the paclitaxel dose and increased mortality. Schneider et al, in a recent publication⁶,demonstrated the same in the IN.PACT SFA cohort of patients. They found no relationship, and in fact, patients who received higher doses actually had lower mortality. Based on the above concerns with this meta-analysis, I question the conclusion that paclitaxel is the cause of increased mortality of these patients.

How will you be using drug-coated balloons at Mount Sinai going forward?

At Mount Sinai and every hospital, and as physicians, we don’t want to do harm to our patients. We are waiting to find out what happens as a result of the FDA panel. We do educate our patients, especially those patients who have repeat procedures and are at high risk for reintervention. The FDA indicates in their letter that repeat intervention, high-risk patients may be appropriate for paclitaxel, because it does reduce restenosis rates. Therefore, if a patient is at a high risk for restenosis, a discussion regarding the use of paclitaxel-coated balloons and stents may be appropriate.

Are there other options?

There is ongoing research into other technologies. The Katsanos meta-analysis showed a correlation without causation with multiple flaws, as discussed above. The danger in this type of conclusion, without a clear identifiable cause, is that it may result in the unnecessary withholding of effective therapy from patients who would otherwise benefit. We need to continue looking at better and newer technologies, such as bioresorbable and alternate drug technology.

What do you think should be happening going forward?

The rigor by which we complete our trials is excellent and will continue to improve. First, mortality may become one of the primary endpoints. Second, we need to evaluate the 2:1 randomization strategy. Third, the studies should have mandated long-term follow-up, and have a mandated registry along with the study that will enroll real-world patients.

Any final thoughts?

Paclitaxel-coated therapy in the superficial femoral artery has been shown to be more effective than PTA in numerous randomized control trials. Any concern over mortality should be addressed based on the quality of the data presented and I await guidance from the FDA regarding this important issue. We will be receiving further data from the Medicare and other data sets so that we can hopefully come to a decision regarding the concern of increased mortality in patients receiving paclitaxel-coated therapy.

Dr. Prakash Krishnan can be contacted at prakash.krishnan@mountsinai.org

REFERENCES

1. Katsanos K, Spiliopoulos S, Kitrou P, Krokidis M, Karnabatidis D. Risk of death following application of paclitaxel-coated balloons and stents in the femoropopliteal artery of the leg: a systematic review and meta-analysis of randomized controlled trials. J Am Heart Assoc. 2018 Dec 18; 7(24): e011245.
2. Treatment of Peripheral Arterial Disease with Paclitaxel-Coated Balloons and Paclitaxel-Eluting Stents Potentially Associated with Increased Mortality - Letter to Health Care Providers. U.S. Food & Drug Administration. January 17, 2019. Available online at https://www.fda.gov/medical-devices/letters-health-care-providers/treatment-peripheral-arterial-disease-paclitaxel-coated-balloons-and-paclitaxel-eluting-stents. Accessed June 25, 2019.
3. UPDATE: Treatment of Peripheral Arterial Disease with Paclitaxel-Coated Balloons and Paclitaxel-Eluting Stents Potentially Associated with Increased Mortality - Letter to Health Care Providers. U.S. Food & Drug Administration. March 15, 2019. Available online at https://www.fda.gov/medical-devices/letters-health-care-providers/update-treatment-peripheral-arterial-disease-paclitaxel-coated-balloons-and-paclitaxel-eluting. Accessed June 25, 2019.
4. Holden A, Varcoe RL, Jaff MR, Schneider PA, Tepe G, Zeller T. Paclitaxel and mortality: the dose argument is critical. J Endovasc Ther. 2019 Jun 9:1526602819857241. doi: 10.1177/1526602819857241. [Epub ahead of print]
5. JEVT editorial challenges drug dosing and exposure time assumptions in paclitaxel mortality meta-analysis. Endovascular Today. June 17, 2019. Available online at https://evtoday.com/2019/06/jevt-editorial-challenges-drug-dosing-and-exposure-time-assumptions-in-paclitaxel-mortality-meta-analysis/. Accessed June 29, 2019.
6. Schneider PA, Laird JR, Doros G, et al. Mortality not correlated with paclitaxel exposure: an independent patient-level meta-analysis of a drug-coated balloon. J Am Coll Cardiol. 2019 May 28; 73(20): 2550-2563. doi: 10.1016/j.jacc.2019.01.013.