Following the publication of the ATTRACT trial in the New England Journal of Medicine (NEJM) in December 2017, Venous News spoke to experts in the field to analyse the data and formulate future research questions. One of the interesting points raised was whether the Villalta scale is an altogether valid measure of post-thrombotic syndrome, even though its use is recommended by international guidelines. The Villalta scale is considered the gold standard for the diagnosis and classification of post-thrombotic syndrome, when combined with a venous disease-specific quality-of-life questionnaire.
The ATTRACT (Acute venous thrombosis: Thrombus removal with adjunctive catheter-directed thrombolysis) triallists set out to assess the treatment effect obtained by adding pharmacomechanical thrombolysis and ancillary measures including thrombus aspiration/maceration, with or without balloon angioplasty/stenting vs. anticoagulation alone on the incidence of post-thrombotic syndrome.
As reported in NEJM, Suresh Vedantham (Professor of Radiology and Surgery, Mallinckrodt Institute of Radiology, Washington University in St Louis, USA) and colleagues randomly assigned 692 patients with acute proximal deep-vein thrombosis to receive either anticoagulation alone or anticoagulation plus pharmacomechanical thrombolysis. The primary outcome of the study was defined as the cumulative incidence of post-thrombotic syndrome, as diagnosed and measured by the Villalta scale, between six and 24 months of follow-up.
Secondary outcome measures included the severity of post-thrombotic syndrome; resolution of presenting deep venous thrombosis symptoms; the prevalence of valvular reflux and residual thrombus, the degree of clot lysis; and cost-effectiveness within 24 months of randomisation. Further, major bleeding, symptomatic pulmonary embolism, recurrent venous thromboembolism, and death within 10 days and 24 months after randomisation formed part of the secondary outcome.
According to the data published in the NEJM, the ATTRACT trial data did not show a significant between-group difference in the percentage of patients with the post-thrombotic syndrome (47% in the pharmacomechanical thrombolysis group and 48% in the control group; p=0.56) between six and 24 months. Pharmacomechanical thrombolysis led to more major bleeding events within 10 days (1.7% vs. 0.3% of patients; p=0.049). However, no significant difference in recurrent venous thromboembolism was seen over the 24-month follow-up period (12% in the pharmacomechanical thrombolysis group and 8% in the control group, p=0.09). “Moderate-to-severe post-thrombotic syndrome occurred in 18% of patients in the pharmacomechanical-thrombolysis group vs. 24% of those in the control group (p=0.04).
Severity scores for the post-thrombotic syndrome were lower in the pharmacomechanical thrombolysis group than in the control group at six, 12, 18 and 24 months of follow-up (p<0.01 for the comparison of the Villalta scores at each time point), but the improvement in quality of life from baseline to 24 months did not differ significantly between the treatment groups,” the authors report.
Based on the data obtained from the study, the triallists concluded that “among patients with acute proximal deep vein thrombosis, the addition of pharmacomechanical catheter-directed thrombolysis to anticoagulation did not result in a lower risk of the post-thrombotic syndrome but did result in a higher risk of major bleeding.”
Stephen Black, consultant vascular surgeon at Guy’s and St Thomas’ Hospital, London, UK, and one of the editors-in-chief of Venous News, says: “The triallists have done an immense job in completing this study to gather these data. While it is easy to be an armchair critic, perhaps the biggest change I would have made is to not have a binary outcome measure of post-thrombotic syndrome.”
“The most interesting ATTRACT trial data were the two tables: one showing binary data (ie., yes or no to the incidence of post-thrombotic syndrome) and the other showing continuous outcome data. What is striking is that the binary data (which was used as the primary outcome measure) shows no difference; however, the continuous data shows that patients who had the more aggressive inte vention were significantly better, based on symptoms at every time point. This suggests that while lysis did not prevent the onset of post-thrombotic syndrome, those patients who had pharmacomechanical catheter-directed thrombolysis were significantly better, based on symptom improvement. In other words, had the endpoint been focused on symptom improvement, this would have been a study that favoured intervention. This appears to have been largely overlooked in the presentation of the trial data.”
Lowell S Kabnick, director, NYU Vein Center, Division of Vascular Surgery, NYU Langone Medical Center, New York, USA, and the immediate past president of the American Venous Forum, notes: “The lack of clinical significance for the primary endpoint between groups was disappointing; however, the decrease in severity of post-thrombotic syndrome with pharmacomechanical catheter-directed thrombolysis is important.”
For Gerry O’Sullivan, vascular and interventional radiologist, University Hospital Galway, Galway, Ireland, the single biggest take-home from the trials was that “regardless of whether you are an interventionalist or an internal medicine specialist, the rate of post-thrombotic syndrome was incredibly high—nearly 50%—in both groups at two years. So whether we use anticoagulation or catheter-directed thrombolysis (or any of the other techniques), at the moment we are not doing these patients any favours as these treatment outcomes are unacceptable.”
Ramesh K Tripathi, adjunct professor, Faculty of Science, Health, Education and Engineering, University of the Sunshine Coast, Australia, states: “ATTRACT showed encouraging trends between the two studied subgroups analysed. Unfortunately, when evaluated separately, they were not sufficiently powered to show statistically significant differences. That may be one of the reasons why this study does not change my current practice protocols. It struck me that only one out of every 52 patients that met the inclusion criteria was recruited into this study. At a sample size of 692 recruited patients, it is assumed that 30% of patients will develop post-thrombotic syndrome during 24 months with a 33% risk reduction in the pharmacomechanical thrombolysis group. In reality, the outcomes at 24 months show a 48.2% post-thrombotic syndrome with anticoagulation compared with 46.7% in the catheter-directed lysis group.”
Is the Villalta scale fit for purpose?
The Villalta scale has widely been used as a scoring system to diagnose and classify post-thrombotic syndrome. International guidelines recommend that post-thrombotic syndrome is assessed using the Villalta scale, a clinical measure that incorporates venous symptoms and signs in the leg ipsilateral to a deep venous thrombosis. In the main, criticism of the Villalta score in the literature appears to be based on its reliance on subjective measures. Arany Soosianathan and colleagues proposed in the Journal of Vascular Surgery (2013) that “the Villalta score, combined with a venous disease-specific quality-of-life questionnaire, should be considered the ‘gold standard’ for the diagnosis and classification of post-thrombotic syndrome.”
Importantly, the post-thrombotic syndrome is a hard entity to precisely define—a “very difficult diagnosis to pin down, particularly in the less severe cases”, according to O’Sullivan.
The ATTRACT trial investigators also alluded to this in the publication. “Because the post-thrombotic syndrome varies in its clinical manifestations, we evaluated its presence and severity in complementary ways. Assessments made with the use of the Villalta scale and the Venous Clinical Severity Score were consistent in suggesting that pharmacomechanical thrombolysis reduced the severity of the post-thrombotic syndrome, which raises the possibility that the aetiologic factors that predispose patients to the development of the post-thrombotic syndrome may differ at least partly from those that determine progression to advanced post-thrombotic syndrome,” the ATTRACT authors write.
O’Sullivan says: “The issue with post-thrombotic syndrome is that there is no method to identify it and measure it, and really the Villalta scale is a rather imperfect method. Indeed, it has never been formally published and therefore, never been subjected to the standard scrutiny that publication brings. Again, this is not in any way to criticise the triallists for employing this; because the Villalta scale is generally accepted as the best measure that we currently have; it is just that it is less than ideal and this is not only problematic for the results of this trial, but also for any studies in the field of deep venous thrombosis research.
“At two years following a deep venous thrombosis, my patients usually do not have ulcers, or skin changes; they might have a degree of oedema, but the most common symptoms are venous claudication and weight gain. There is currently no accepted measure for venous claudication, which is incredible. In the arteries we have the ankle brachial pressure index (ABPI; a method for quantifying the severity of arterial occlusion in the leg), yet there is no such comparable measure in veins and this is a huge problem.
“The bigger issue with the Villalta scale is that it has been shown fairly conclusively (by Fedor Lurie et al, for instance) that even without a history of deep vein thrombosis, a patient can achieve a score of five on the scale and by the definition employed in ATTRACT be labelled as having post-thrombotic syndrome. So it is questionable to me at least what the nearly 50% rate really means.”
Black echoes these doubts about the Villalta scale as a reliable measure of post-thrombotic syndrome. “One key question for me remains the outcome measure we choose to assess these patients. Villalta is undoubtedly inadequate and until we find a more robust measure to assess these patients, we will continue to find it difficult to accurately answer questions about post-thrombotic syndrome,” he notes.
Inclusion of femoropopliteal patients
Many experts also believe that the inclusion of femoropopliteal patients into the trial might have confounded its results. Since recruitment into the trial was slow, the investigators pursued a strategy of including patients with femoropopliteal disease as well as iliofemoral disease (56% of the cohort).
Kabnick commented that with designing a trial such as ATTRACT “there are always trade-offs”. “The cohort of patients with acute femoral-popliteal deep venous thrombosis could have confounded the results of the trial. Thus, the primary outcome measure may have been attainable by decreasing the magnitude of post-thrombotic syndrome through pharmacomechanical catheter-directed thrombolysis in iliofemoral deep venous thrombosis. However, without the addition of the femoropopliteal segment, completion of the decade-long study would have been longer and other study obligations would have surfaced. Regardless of these results, or lack of results, I believe that clinicians will strongly consider procedural intervention in the patients with acute iliofemoral deep venous thrombosis. Of course, further investigation is necessary to determine treatment as well as eligible candidates.”
Tripathi states: “The sample pool was contaminated by the inclusion of more distal deep venous thrombosis cases (43% overall) in order to ensure adequate enrolment.”
Black agrees: “Clearly the slow rate of enrolment hindered the study and led to the inclusion of femoral deep venous thrombosis patients which has hampered the analysis.”
Venous stenting following thrombus removal
Tripathi also commented on the low stenting rate in the trial. “Although 62% patients underwent iliac venoplasty, only 28% underwent iliac vein stenting. Given the high rate of recoil of lesions that underwent venoplasty, this may be of critical importance as in real life practice most of these patients would undergo stenting,” he said.
O’Sullivan talks about the low stenting rate in the trial by referring to current European thinking that 80–90% of iliofemoral deep venous thrombosis patients, following removal of thrombus “would typically need a stent.” He also commented that future research would need to mandate intravascular ultrasound (IVUS) and multiplanar venography so that the procedural imaging protocols reflect current practice.
Tripathi concurs: “In this study, IVUS was not used to assess luminal diameters or adequacy of deep venous thrombosis clearance and subsequent endovascular treatments towards achieving the concept of “open vein hypothesis”. Instead modified Marder Scores were used, that are arbitrary, based on subjective observations of clot volume and therefore open to inaccuracies of interpretation. It is also limited by its failure to separate the relative contributions of thrombus within the iliac veins, and the contributions of the profunda femoral vein and the inferior vena cava. I would like intravascular ultrasound to be part of any future study to objectively assess post–thrombolysis/post-venoplasty evaluation of residual stenosis,” he says.
So which patients are suitable for pharmacomechanical lysis?
Experts opine that the ATTRACT trial is a really important addition to the literature as it showed “encouraging trends” in the subgroups analyses.
ATTRACT data from subgroups and secondary analyses suggest that catheter-directed thrombolysis may have a benefit in patients who have acute iliofemoral deep vein thrombosis. The data also showed a trend towards greater treatment effect for catheter-based interventions with patients who present with more severe symptoms.
However, triallists emphasise that ATTRACT was not statistically powered to draw definitive conclusions from these secondary analyses and outcomes and that the treatment effect seen in these should be seen as trends only.
Commenting on the patients for whom pharmacomechanical thrombolysis would be a suitable treatment option in light of the trial data, Black notes: “ATTRACT suggests we should not be offering lysis except for severe patients, so it is difficult to answer this based on the presented data. My view is that we will continue to offer lysis to patients with iliofemoral deep venous thrombosis, particularly to those who have significant symptoms (pain and swelling plus venous claudication) on presentation.”
O’Sullivan suggests: “The ATTRACT trial showed us who not to treat, and therefore also who we should potentially concentrate on, namely, iliofemoral patients under the age of 70, who are very symptomatic. This is what we are effectively doing now and have done so for some time. I cannot recall the last time I treated a patient with purely femoropopliteal disease.”
Tripathi agrees: “A young patient (age <65years) who has iliofemoral deep venous thrombosis with severe symptoms and a low risk of bleeding will most likely benefit in terms of reduction of acute deep venous thrombosis symptoms and post-thrombotic syndrome severity.”
Kabnick believes: “Catheter-based procedures will be offered to the appropriate patient; however, this should not be based on age! For those patients who are not candidates for procedural intervention, the ATTRACT trial in of itself has justified anticoagulation.”
The experts agree that future research needs to focus on iliofemoral disease as the treatment group most likely to benefit from catheter-based therapies, so as to usher in a more targeted approach to therapy. It will hopefully also redirect efforts to better mechanisms of clot removal and care of the residual vein. They say imaging follow-up should involve the use of colour Doppler ultrasound (used in about 200 of the nearly 700 patients in the ATTRACT trial, which meant that in a large proportion of patients, it was unknown whether veins were open or blocked).
There should also be improvement in patient selection; better definition of the extent of disease and an improved measure of outcomes, they add.