Ultrasound-assisted catheter directed therapy (CDT) for pulmonary embolism versus standard CDT: Sounds of a cry for data!

Recently, there has been a tremendous surge of energy to diagnose and treat submassive and massive pulmonary embolism (PE). This movement has prompted many health care systems to develop PE Response Teams (PERT), often composed of multiple specialties with expertise in the management of this potentially life-threatening condition. Though historically treated with anticoagulation alone, and/or systemic thrombolysis or open surgical pulmonary embolectomy, catheter-directed thrombolytic therapy (CDT) has evolved into a potential treatment option ¹ for massive and submassive PE. The American Heart Association endorses CDT as a viable alternative to systemic thrombolysis, ² highlighting its role in the PE armamentarium. Considering that 45–50% of patients with PE present with signs and symptoms of right ventricular strain (submassive) and/or hemodynamic instability, including cardiovascular collapse (massive), the public health impact of expanded treatment options for this condition cannot be underestimated. As such, there has been a growing interest in treating PE using CDT to potentially prevent chronic thromboembolic pulmonary hypertension (CTEPH) and minimize complications which have been associated with systemic lysis, especially bleeding.

There are two distinct types of CDT: one using a standard infusion catheter placed into the pulmonary arteries through which thrombolytics are administered, and the other that uses ultrasound to assist in the thrombolysis (pharmacomechanical thrombolysis). The latter (i.e. the EkoSonic Endovascular System; recently acquired by Boston Scientific, Natick, MA, USA) uses ultrasonic energy to purportedly disrupt fibrin and expose more binding sites for the thrombolytic agent, ³ thereby potentially dissolving more of the clot faster. One early, nonrandomized study, ⁴ showed better thrombus removal with reduced thrombolytic infusion time and reduced treatment-related complications compared to standard CDT. Since then, use of this technology (i.e. ultrasound-assisted catheter-directed therapy (UA-CDT)) has exploded, especially in the face of clinical trials Ultima ⁵ and SEATTLE II ⁶ demonstrating its efficacy and safety profile compared to anticoagulation alone in patients with submassive PE. These catheters, however, are markedly more costly than the standard catheters used for CDT. Moreover, there are no prospective head-to-head data demonstrating a benefit of UA-CDT over CDT alone.

In this issue of Vascular Medicine, both Rao and colleagues ⁷ and Rothschild and colleagues ⁸ present retrospective data supporting the clinical equipoise for UA-CDT and CDT from two different, highly active PERT centers. These studies highlight the importance of comparative effectiveness trials in the era of value-based care and raise important questions about the clinical utility of UA-CDT over CDT alone.

Both studies retrospectively evaluated the clinical efficacy and safety of UA-CDT versus CDT for the treatment of submassive PE at each center. Rothschild and colleagues amalgamated the data at Beaumont Hospital over the last 8 years, which included 98 cases (62 UA-CDT and 36 CDT), and examined efficacy and safety data between the two cohorts. Specifically, they examined the impact of the clot on the right ventricle (RV) with the change delta in pulmonary artery (PA) systolic pressure over 24 hours, the RV/LV diameter ratio, and general clinical outcomes such as length of stay, bleeding, and mortality. They report essentially no difference between the two treatment arms and suggest that the more expensive UA-CDT may not be required to achieve similar outcomes. Though the authors made an attempt to adjudicate for bias and confounding, there are other issues with this study that must be considered. The small number of patients included over an 8-year period leads to a dynamic cohort treated with a variety of diverse techniques and lytic doses/duration, distorting how to interpret the data. Furthermore, at this institution, low-dose systemic thrombolysis was often used (rather than CDT), creating a selection bias. The lack of uniformity across these domains weakens the conclusions of the study. Furthermore, the study may lack power to see small differences.

Rao and colleagues ⁷ also retrospectively examined the same question (UA-CDT vs CDT) using data from 70 patients with PE who presented to four centers within the same health care system over a 3-year period (2014–2017). Of the 70 patients, 37 were treated with UA-CDT and 33 received standard CDT. In addition to the clinical outcomes, these investigators attempted to further examine the quality of life outcomes of these patients. They also report no difference in their clinical outcome variable of reduction in RV systolic pressure (RVSP) after the procedure, and no difference in bleeding or mortality. Given the more contemporary time period, the two groups in this study were more similar in their treatment time and lytic dose, though the UA-CDT group tended to receive a higher total dose of thrombolytics. Interestingly, regardless of the type of CDT, 60–65% of patients still exhibited signs of RV dysfunction as assessed by echocardiography at 24–48 hours. The authors should be congratulated for examining quality of life outcome in patients with PE; however, this was accomplished via telephone 1–4 years after their treatment. Quality of life scores, when assessed after such a significant time delay, as in this case, are prone to recall bias and may be unreliable.

Though important, both studies are hypothesis-generating and are limited by their non-randomized, retrospective nature and small sample size. They join the ranks of several other studies that have drawn similar conclusions both with PE ⁹ and iliofemoral deep vein thrombosis (DVT).^10,11 Taken together, these studies send an important signal of clinical equipoise between the two types of CDT. More importantly, in the case of PE, they highlight a larger problem, a lack of prospective, randomized data showing what indeed is the best type of CDT.

In a recent review of the evidence strength behind the American Heart Association/American College of Cardiology (AHA/ACC) endovascular and surgical guidelines for the treatment of peripheral vascular diseases by Sardar et al., ¹² the lowest level of evidence for the statements were observed for PE/DVT interventions (Level of Evidence categories: A = 0%, B = 24%, C = 76%). This underscores the paucity of data in this vascular space, partly because of a lack of consensus on endpoints and the financial disinterest to conduct comparative effectiveness studies. In the two retrospective studies in this issue, a reduction in RVSP or pulmonary artery systolic pressure (PASP) was used as an efficacy outcome; however, in the ULTIMA trial ⁵ (UA-CDT vs anticoagulation alone), the outcome variable was the RV/LV ratio at 24 hours. In the SEATTLE II trial ⁶ (150 patients treated with UA-CDT alone with no comparison), the mean RV/LV ratio from baseline to 48 hours, mean PASP, and modified Miller Index score (measure of clot burden on computed tomography (CT) angiogram) were reported as outcomes. Collectively, this lack of standardization on the appropriate PE outcome, timing as to when it should be measured, and even methods of measurement, makes any comparisons difficult.

Beyond the mode of therapy, the more fundamental question for this field is: what outcome should we be measuring in PE trials? Early studies of PE showed that most patients with PE treated with anticoagulation alone achieve embolus resolution at 4 weeks, though most of these would have been low risk PE cases.^13,14 In addition to what outcome is optimal to be measured, when it is measured is just as important. Older studies of patients with PE treated with anticoagulation alone, demonstrated (via serial pulmonary angiograms and perfusion lung scans) that the resolution of PE is negligible after 2 hours, ~10% after 24 hours, 40% after 7 days, and 50% after 2–4 weeks, whereas thrombolytic therapy only accelerates the rate of resolution (i.e. to ~10% at 2 hours, 30% at 24 hours, 45% at 7 days, and 50% at 2–4 weeks) but does not alter the extent of residual thrombosis. ¹³ These natural history data make us question the outcomes of 24-hour RV/LV ratios, clot burden scores such as the Miller index, or any of the outcomes followed by the UA-CDT trials. What is not known is how such a trial today in a population of patients with true high-risk submassive PE (positive biomarkers and RV dysfunction as assessed by echocardiography) would pan out. In fact, many CDT antagonists cite these older natural history studies of PE as why they choose anticoagulation alone for most cases. This negative argument is further strengthened by actual data from the ULTIMA trial, which showed no difference in the RV/LV ratio comparing those treated by UA-CDT versus anticoagulation alone at 90 days. ⁵

The cry for prospective, randomized trial data in this arena is rising to a deafening din as the demands for data-driven, cost-effective practice only become louder. Though the ongoing randomized multicenter SUNSET sPE ¹⁵ (NCT02758574) trial of UA-CDT versus CDT alone will further elucidate whether UA-CDT is cost-effective, it too will be limited in its relatively small numbers, selection bias, and at least some variation in thrombolytic dosing and/or duration.