Stavros V. Konstantinides, MD
Professor, Clinical Trials
Medical Director
Center for Thrombosis and Hemostasis
University Medical Center Mainz
Mainz, Germany
stavros.konstantinides@unimedizin-mainz.de

Suresh Vedantham, MD
Professor of Radiology and Surgery Mallinckrodt Institute of Radiology
Assistant Dean for Clinical Research
Washington University School of Medicine
St. Louis, Missouri
vedanthams@mir.wustl.edu

First, how would you describe the current landscape of clinical study in the field of pulmonary embolism (PE) intervention?

Dr. Vedantham: It’s an exciting time, with great potential to advance the standard of PE care. I’m glad to see broad recognition that we need quality evidence to direct care and investments to match. It is now 15 years after the United States Surgeon General issued a National Call to Action on venous thromboembolism (VTE), citing the “Power of Partnerships.” It is great to see the multidisciplinary collaborative energy continue.

Prof. Konstantinides: The landscape is rapidly evolving. We had an almost 10-year period in which valuable observational, or prospective but noncontrolled, data were obtained. These data provided important information on the efficacy and safety of catheter-directed treatment (CDT), but it could not, by design, convincingly “prove” its clinical benefits. Accordingly, CDT remained a topic of “belief” rather than science, but this is changing. Now, a number of high-quality, randomized controlled trials (RCTs) with adequate sample size and clinical primary endpoints are underway. These will inform future guidelines recommendations and clinical practice in the following 2 to 3 years.

How would you compare and contrast this with the study of deep vein thrombosis (DVT) interventions?

Prof. Konstantinides: The situation with CDT in DVT has thus far been quite similar to that outlined previously. Our advantage in the PE field is that we can directly demonstrate acute clinical benefits if we appropriately design our trials. This is what we are doing right now. DVT trials cannot do this because it is not an immediately life-threatening situation; however, what we learned from DVT trials is to also focus on long-term outcomes and patient-related outcome measures. In PE, the postthrombotic syndrome (PTS) is not an outcome, but we do have “post-PE syndrome,” which encompasses a broad range of clinical symptoms and functional abnormalities as well as generic and disease-specific quality-of-life (QoL) indicators. We are concentrating on all these outcomes in ongoing and planned PE trials. So, both fields can learn from each other.

Dr. Vedantham: In different ways, the environment is similar to 15 years ago when the National Institutes of Health (NIH)–sponsored ATTRACT trial was starting. New devices were available, prompting excitement, but there were no completed trials of high methodologic quality. Clinical change over time was automatically attributed to clot removal, reflecting wishful thinking and unfamiliarity with the natural history of DVT treated with anticoagulation alone. Through RCTs, we learned that much of this clinical change was simply due to standard care, not the procedures. This may or may not prove similar for PE, but it is good that we are finally trying to find out!

Clinical study of PE may prove easier than DVT because (1) severe PE is an inpatient condition—in hospital systems, its importance is apparent to many stakeholders; and (2) the collaborative clinical development around PE has been strong. Specifically, the formation of multidisciplinary PE response teams (PERTs) around the country has engendered the intense multispecialty engagement that is important in designing studies and completing them in a timely fashion.

How has experience elsewhere in VTE trials, such as ATTRACT in DVT, informed the unique needs in designing future PE trials?

Dr. Vedantham: The NIH-sponsored PE-TRACT trial, led by Principal Investigator Dr. Akhilesh Sista, arrives in direct lineage from ATTRACT and C-TRACT. The structure of its multidisciplinary steering committee and investigator network, consensus-based protocol development, scientific rigor, and community engagement were heavily informed by these previous NIH studies. In particular, the strong commitment to giving endovascular PE therapy the “best fair chance” to succeed, allowing different treatment methods, assessing patient outcomes independently, minimizing bias, and reporting the results with full transparency are similar to ATTRACT and C-TRACT.

I believe PE studies have also benefited from the challenges faced by ATTRACT and C-TRACT. I think ATTRACT was quite forward-looking in its use of informatics tools, our constant proactive education of the site network on best practices for participant recruitment, and our willingness to regularly reinvent our approach. Having said that, ATTRACT took twice as long as projected to complete accrual, and C-TRACT is suffering similar tribulations. It is great to see that ongoing PE studies have recognized this challenge and are using modern tools (artificial intelligence, electronic informed consent with video aids, others) to overcome it. PE-TRACT also has an adaptive design that permits trial completion with fewer patients if justified. With the support of the PERTs, I expect PE trials to be completed in a more timely fashion. I’m encouraged that PE-TRACT has gotten out of the enrollment gate quickly in its early months.

Prof. Konstantinides: As I just emphasized, we have learned from the DVT trials to turn our attention to long-term clinical and functional outcomes and to patients’ well-being and QoL. We have also learned that we must clearly describe, standardize, and harmonize (amongst sites) the CDT procedure in the trial protocol in order to have results that can be extrapolated to all centers performing these procedures or wishing to do so in the future. It is no longer sufficient to design a trial in which a vague term such as “catheter thrombolysis” (of any dose and any duration) is compared to another therapeutic modality.

What are the key lessons learned from the single-arm PE studies that have been conducted to date? What are their strengths and limitations?

Prof. Konstantinides: These studies have shown that CDT works in terms of relieving pressure overload of the right ventricle and reducing its size (ie, improving its function). They have also shown that CDT is generally safe in terms of major and, most of all, intracranial bleeding complications, and most likely safer than systemic thrombolysis. However, what these studies could not show is if these results directly affect patient prognosis compared to the current standard of care, which for intermediate-risk PE at least is “simply” anticoagulation. This can only be done in a head-to-head comparison with the current standard of care in an appropriately designed, adequately powered, multicenter (preferably multinational) RCT.

Dr. Vedantham: The single-arm PE studies have shown that it is feasible in academic medical centers and community hospitals to manage sick PE patients in endovascular procedure suites, position catheters/devices in the pulmonary arteries (PAs), and perform interventions that reduce thrombus burden and PA pressures. In massive PE, the rapid timing of clinical change suggests a positive effect of the intervention. However, complications can occur. With ultrasound-assisted catheter-directed thrombolysis (CDL), major bleeding in the SEATTLE II registry exceeded 10%; two patients in the OPTALYSE study had intracranial bleeds. With mechanical thrombectomy (MT), there have also been complications, some of which were likely procedure-related.

The most significant limitation of the single-arm studies is that they cannot determine the extent to which any observed clinical change can be reasonably attributed to the catheter-based intervention versus standard medical therapy, close monitoring, and natural healing. Because these interventions pose risks and costs, this crucial question absolutely must be answered!

There is good reason to take a skeptical, “study-first” posture. In the PEITHO trial (which evaluated systemic thrombolysis), 95% of patients with acute intermediate-risk PE did not deteriorate, and > 98% of patients survived during the first 7 days with standard care alone. In the OPTALYSE study, a right ventricular/left ventricular (RV/LV) ratio reduction was seen 48 hours after CDL in all treatment arms, including the lowest recombinant tissue plasminogen activator dose regimen in which thrombus reduction was minimal (5% reduction in modified Miller index). This suggests that part or all of the observed RV/LV ratio change may simply result from standard care and time, not clot removal. Logically, that is not inconsistent with the observation that most patients improve after 48 hours of anticoagulation. Only through rigorous RCTs trials will we learn if patients truly benefit from intervention.

The field has responded to demands for more randomized data with a handful of new industry-backed trials. What are the strengths and limitations of this new generation of PE trials in your opinion? What will they add to the body of evidence?

Prof. Konstantinides: Yes, the field is now responding to the needs outlined previously. It is great news that the industry has endorsed the concept of convincingly demonstrating the clinical benefits of CDT. Industry-sponsored trials are now underway, and importantly, the industry is closely collaborating in some of them with academic institutions and not-for-profit organizations (eg, in the HI-PEITHO trial, Boston Scientific Corporation [the trial sponsor] is conducting the trial in partnership with The PERT Consortium in the United States and the University Medical Center of Mainz in Germany). These trials are not only methodologically state-of-the-art but also have the “right” clinical primary outcome and, in addition, the long-term (over at least 1-year follow-up) outcomes described previously: “post-PE syndrome” with a broad range of clinical complaints and functional abnormalities that are being evaluated at follow-up visits, and generic and disease-specific QoL indicators that are being assessed with standardized questionnaires. These outcomes have the strength and validity to convince clinicians, guidelines task forces, and policymakers (for CDT reimbursement) in the United States, Europe, and all over the world, depending on the upcoming results of these trials. CDT will soon be a matter of scientific evidence instead of the fight between believers and nonbelievers.

Dr. Vedantham: I am glad to see randomized studies in the endovascular PE therapy space, and I like some elements of all of them. The holistic stance they take toward characterizing PE outcomes is mostly a strength, in my view. I like the application of creative strategies (eg, hierarchical win-ratio methods that prioritize outcomes of greatest impact) in some of them to obtain insight into the different dimensions of PE outcomes that are important to patients and clinicians. After these trials are completed, the meaning and value of some of the surrogate outcomes (eg, RV/LV ratio) should become clearer via correlation with other clinically relevant outcomes.

Having said that, I presume that sponsors and investigators harbor hopes that these studies will transform PE care and flip clinical practice guidelines. Speaking candidly, some aspects of these studies reduce my confidence that they have minimized bias sufficiently, certainly in the eyes of the most rigorous and relevant arbiters. Randomization addresses patient selection bias, but finding truth hinges on systematically minimizing as many potential sources of bias as possible. That is what high-impact journals and rigorous guideline panels (eg, CHEST, American Society of Hematology) will expect to see before favoring recommendations for catheter intervention.

First, RCTs should compare populations that can reasonably receive either treatment. A study comparing CDL and MT should exclude patients with major contraindications to either therapy. If patients who have had recent surgery or other major contraindications to fibrinolytic drugs are included, it can artificially bias the safety comparison. Second, clinical and imaging outcomes need to be assessed and adjudicated by independent, blinded, well-qualified observers, using processes and precautions that are standardized and explicitly designed to reduce bias. Third, in designing composite or hierarchical outcome assessment plans, care should be exercised in incorporating subjectively determined outcomes and in mixing clinical and economic/efficiency outcomes. Fourth, and probably of greatest concern, if major outcomes can be declared as having been met by involved (unblinded) physicians based on clinical deterioration or “need for bailout procedure,” this can create bias. In fact, this type of deficiency was heavily criticized in prior PE studies that evaluated systemic thrombolysis. Unplanned endovascular procedures may occasionally be needed in clinically worsening PE patients who are enrolled in RCTs, but these occurrences do not all need to trigger endpoint determinations.

In this respect, I respectfully challenge all physician investigators to be stronger advocates for patients and quality science. We should collaborate with industry and value their important role in driving forward progress. But leadership is not just “going along” and enrolling patients in studies with obvious sources of bias. Leadership is strongly advocating for what gets accurate and unbiased answers for patients, putting their rights and safety first, and embracing the fact that physicians actually have the strongest influence in what patients and sponsors choose to do. No PE treatment trial can enroll a patient without a supportive physician. Before studies start, investigators should share concerns about study design with each other and make sure they are addressed by study leaders and sponsors. Sometimes these conversations will be difficult, but they need to occur nonetheless. We will only find truth and rapidly advance patient care if we use our voices collectively and insist on rigorous studies with unbiased assessments.

PE-TRACT has also gained NIH funding and begun enrollment. Dr. Vedantham, can you give us a brief summary of this trial’s design and potential impact?

Dr. Vedantham: PE-TRACT is a multicenter RCT involving 30 to 40 United States–based hospital sites. Patients with acute intermediate-risk PE with central clot burden and an RV/LV ratio exceeding 1.0 are being randomized to receive CDT (which can include CDL or MT using devices that have been cleared by the FDA for PA use) or no CDT. All patients receive anticoagulant therapy and close monitoring. PE-TRACT is unique in being designed to evaluate midterm outcomes using an innovative adaptive design strategy. Specifically, the study has two primary outcomes: (1) objective evaluation of peak oxygen consumption on cardiopulmonary exercise testing at 3 months postrandomization and (2) patient-reported New York Heart Association (NYHA) class at 12 months postrandomization. To provide confidence that a difference in 12-month NYHA class is truly related to the PE intervention (and not comorbidities unrelated to PE), the two outcomes are linked by a gatekeeping strategy whereby statistical testing for benefit on the NYHA outcome will only occur if statistical testing of the objective 3-month oxygen consumption is positive. Prespecified interim analyses will be conducted starting at 250 patients and then every 50 patients, up to a maximum of 500 patients.

In addition to these primary outcomes, PE-TRACT will evaluate early outcomes such as clinical deterioration and safety, as well as bleeding, recurrent VTE, and health-related QoL. Importantly, the study has robust precautions to prevent bias in the assessment of clinical outcomes. The collection of clinical, physiologic, and biospecimens data in the context of a rigorous NIH randomized trial gives PE-TRACT a unique ability to provide biological insights into the predictors and mechanisms of post-PE syndrome and which patients benefit from CDT.

This is important since clinical progress is rarely achieved through one study but rather by sequential studies reflecting refinement of study hypotheses with the benefit of the studies before. Like ATTRACT before it, PE-TRACT stands a strong chance of impacting clinical practice guidelines and will create a one-of-a-kind platform for future progress in the field. I sincerely hope that the PE treatment community prioritizes its support for this pivotal clinical study.

While we are on the topic of DVT, what are the most pressing questions to be explored in future studies?

Dr. Vedantham: Most of all, we need to understand the incremental benefit and risk of using MT as a first-line treatment for acute iliofemoral DVT. Many practitioners are proceeding as if the benefits seen with pharmacomechanical CDL in ATTRACT’s iliofemoral subgroup can be applied to MT, but in truth, these are very different treatments. At present, there are no completed MT studies with independent assessments, so there is strong potential for bias in the published data. After enormous effort from many parties, the ATTRACT, CAVENT, and CAVA trials elevated our standards of rigor in acute DVT studies, and I will continue to advocate strongly for RCTs evaluating MT to apply similarly high standards—why would we want to go backward? I have expressed concerns that in an ongoing multicenter RCT evaluating MT, bias can be easily introduced from the lack of blinding in clinical outcome assessment and from the ability of physicians to unjustifiably declare “treatment failure” in control arm patients. I have heard that the protocol may evolve, and I hope so—my concerns will continue until these elements change. I and hopefully others will continue to be vocal in calling out deficiencies in study design, and I look forward to providing even louder support for studies that are rigorously designed.

Certainly, we want to understand the effects of MT on PTS (long term), early symptom improvement, and health-related QoL through a well-designed RCT. But at a much more basic level, simple pilot prospective studies do not exist that characterize the amount of thrombus that is removed with different MT devices in an unpolluted way. Existing studies lack independent, blinded outcome assessments (clinical and ultrasound) and venograms that quantify the effects of thrombectomy alone (before stenting).

We also need to learn the best antithrombotic strategies for the peri-intervention period and understand the role of inflammation more broadly. Anti-inflammatory strategies are being evaluated in clinical trials for their ability to prevent or reduce PTS—this includes rosuvastatin, diosmin, and periadventitial steroid administration. We need to reach beyond that to also understand the effects of catheter intervention on the biology of the venous wall, valves, and blood components. It should be sobering that despite successful initial thrombus removal in the intervention arms of the CAVENT and ATTRACT studies, a substantial proportion of treated vein segments were noncompressible on later imaging follow-up. So, it is possible that the efforts and skill of endovascular proceduralists are being undermined by an unhelpful biological response. We need to understand such effects and study how to overcome them if they exist.

How has the field’s understanding of optimal endpoints in PE trials changed in recent years? What do you feel are the most important outcomes to be examined? Which have come to be considered less so, and why?

Dr. Vedantham: I’ll start by saying that my belief is that most of the PE outcomes studied to date—both those that assess “surrogate” measures and those that assess clinical outcomes with patient-relevant impact—can provide worthwhile insight into the disease processes and the effects of catheter intervention. A lot of focus and debate has been placed on critiquing which outcomes to study. To me, the problem is not which outcome but rather how we are assessing the outcomes and how we are training physicians to recognize bias and confounding. For example, the RV/LV ratio probably does have some relevance to the pace of cardiac recovery after PE. The problem is not use of the RV/LV ratio per se; it is that physicians are attributing the change to an intervention without considering the possibility that the change is simply due to standard care alone.

But to your question, it is good to see a transition toward evaluating clinically important major outcomes. I am glad that major clinical events, objective physiologic parameters, and patient-reported well-being are all being assessed in ongoing trials because they are all important and complementary. I believe PE-TRACT marks a high water in terms of the innovation that has been applied to PE outcomes assessment. For many years, review panels insisted on mortality as the only possible primary outcome for a PE trial, a significant barrier to progress. However, the PE-TRACT investigators collaborated with outstanding biostatisticians and were ultimately able to convince the NIH that (1) it is legitimate to study long-term outcome (post-PE syndrome) as the primary outcome in a PE trial; (2) the aforementioned gatekeeping strategy achieves an overall assessment that is objective and clinically relevant; and (3) the study’s collection of important physiologic and biological data will enhance future PE knowledge generation.

Overall, it’s amazing to think that in a few years, we will have data from multiple large RCTs to guide PE clinical practice, research directions, and future innovation. That’s how the needle was moved for acute stroke care, and I’m optimistic that it will prove the same for patients with PE.

Disclosures

Prof. Konstantinides: Received personal lecture/advisory fees and research grants via his institution from Bayer AG, Boston Scientific, Daiichi-Sankyo, LumiraDx, and Penumbra Inc.

Dr. Vedantham: In-kind research support from Medi USA (donation of compression stockings to study patients; no funding); receives research grants from the National Institutes of Health/National Heart, Lung, and Blood Institute (NIH/NHLBI) for roles as National Principal Investigator of the C-TRACT trial (UH3HL138325) and National Co-Principal Investigator of the PE-TRACT trial (4UH3HL155798). Comments are solely from the author and do not represent the views of the NIH/NHLBI.