Response to Gwozdz and colleagues

Gwozdz and colleagues raise important points about the ATTRACT trial that allow us to put our results into context.^1,2 They reach for the conclusion that the pharmacomechanical catheter-directed thrombolysis (PCDT)-mediated thrombus removal and valvular reflux findings in ATTRACT do not reflect real-world practice, but we respectfully disagree. In particular, caution must be exercised in comparing results across different imaging modalities, time points, and definitions of venous patency.

Regarding thrombus removal, was technical success truly as low as they suggest? In their letter, Gwozdz et al. cite data from the 1-month ultrasound assessments of venous compressibility and residual thrombus that we summarize in our current Vascular Medicine paper to comment on PCDT’s ‘technical success’. However, the technical success of the PCDT procedure is best reflected by the immediate post-PCDT venograms on procedure day. In our original report on ATTRACT, ³ we stated that the mean thrombus score was just 2.7 points out of 24 possible points on the post-PCDT venograms (graded in an independent core laboratory). We can add that the mean post-PCDT thrombus score for the iliac and common femoral vein together was just 0.7 points out of 10 possible points. The operators reported spontaneous anterograde flow in the iliofemoral and femoral-popliteal venous segments on 98% and 88% of the final venograms, respectively. Hence, the immediate PCDT result in ATTRACT was similar to what has been seen in previous prospective studies.^4–8

We agree that the amount of thrombus at 1-month follow-up appears substantial and could have been a factor in the ineffectiveness of PCDT for prevention of reflux and post-thrombotic syndrome (PTS). However, to be called ‘compressible’ in ATTRACT, a vein had to be 100% compressible throughout its length – a more stringent standard than other ‘patency’ definitions. In essence, veins with any residual thrombus (many of which would be called ‘patent’ in other studies) were called ‘non-compressible’ in ATTRACT. This method had the advantage of being amenable to consistent adjudication but must be kept in mind when comparing findings across studies.

Regarding valvular reflux, Gwozdz et al. cite a previous single-center study of infusion catheter-directed thrombolysis (CDT) ⁹ as a standard against which to measure our findings. However, single-center studies often overestimate treatment effects compared to what is seen in well-controlled multicenter trials. ATTRACT was a large, rigorously designed, randomized trial sponsored by the National Institutes of Health with robust precautions against bias and core laboratory assessment of imaging studies. In contrast, the study Gwozdz et al. cite reports very low rates of valvular reflux but did not use an independent core laboratory. ATTRACT employed guideline-endorsed delivery of PCDT and anticoagulant therapy, as well as rigorous pre-study investigator credentialing, source document verification of data, and systematic quality control. Because of its breadth (patients enrolled in 56 diverse clinical centers) and unprecedented rigor, we contend that the results of ATTRACT are more likely to reflect real-world clinical practice than any single-site study.

A better comparator would be the CAVENT trial, a rigorously conducted multicenter randomized trial that evaluated infusion CDT for PTS prevention. ⁶ Like ATTRACT, CAVENT observed good initial thombus removal. However, the immediate post-CDT thrombus score did not correlate with the development of PTS. ⁸ Moreover, substantial residual thrombus and valvular reflux were noted on ultrasound in the CDT-treated patients – at 6 months, only 66% had patent veins and just 29% were free of residual thrombus and valvular reflux. ⁷ These findings are generally consistent with our observations in the current ATTRACT ultrasound study. Also, although CAVENT did observe treatment arm differences, the effect of CDT was underwhelming – approximately a 15% absolute risk reduction for venous patency and reflux, which translated into a small reduction in PTS and no effect on long-term quality of life.^6,10

This raises a few key points. First, while PTS may partly depend on thrombus removal, there are clearly other mechanisms at play. Second, although recurrent symptomatic deep vein thrombosis (DVT) has been infrequent after CDT or PCDT, subclinical thrombus formation may be more substantial than previously appreciated. Unfortunately, no quality studies have prospectively compared different antithrombotic therapy options in patients who have recently undergone endovascular intervention. Third, because the various thrombus removal methods differ in the duration of exposure to fibrinolytic drugs and the extent of catheter manipulations, they may exert different effects upon thrombus reformation, inflammation, and other determinants of venous injury.

We acknowledge the limitations conferred by the modest size of our study. However, it is among the largest studies to have systematically evaluated imaging findings after endovascular DVT therapy, and its methodological rigor should increase confidence in our findings.

The willingness to base clinical practice on the results of well-designed randomized trials is relatively new for the venous endovascular community. As illustrated in the letter by Gwozdz et al., some physicians continue to cite a preference for clinical impressions and studies of less optimal design. All information has value, but we remain convinced that patients will be best served when physicians assign greater value to the minimization of bias that is achieved in randomized trials, and use these data to question assumptions about what is truly needed to improve patient outcomes. The data in ATTRACT suggest that the open vein hypothesis is indeed worthy of continued study. But after unimpressive results for CDT/PCDT in three well-designed randomized trials,^1,6,11 it is time to move beyond overly simplistic formulations of the relationships between thrombus removal and PTS. Yes, clinical outcomes benefit from reduction of thrombus burden – but only to a certain degree, and additional efforts are needed to understand how this can be safely achieved and sustained over time in patients with DVT. In this respect, we believe ATTRACT and the ultrasound analysis presented in Vascular Medicine have significantly advanced our knowledge of therapy for acute DVT.