Acute Deep Vein Thrombosis: Selecting the Right Patient for Thrombolysis

It seems like common sense that patients with deep vein thrombosis (DVT) of the extremities would experience fewer long-term complications (post-thrombotic syndrome, or PTS) if they underwent prompt clot removal. Indeed, a 2016 Cochrane review of 17 randomized trials on the benefit of thrombolysis in lower extremity DVT to reduce PTS found that thrombolysis increased vein patency and reduced incidence of PTS by one-third with a “small increased risk of bleeding.”¹  However, the evidence was judged to be of only moderate quality given the small size of the trials — nine studies had fewer than 50 participants, and only two studies had more than 100. Subsequently, a larger yet still relatively small study (CaVenT trial) of 209 patients who were randomized to catheter-directed thrombolysis or no intervention showed that at two years there was less PTS in the intervention group — 41 percent versus 56 percent (p=0.047). However, this was at the cost of several major and clinically relevant bleeds.²  The benefit in respect to a lower incidence of PTS was maintained at five years; however, there was no associated benefit to patients’ quality of life.³ 

Given this background, a larger and more definitive trial was initiated. The aim: describe the gross clinical benefit of catheter-directed thrombolysis. An additional perk of such a study would be to inform resource utilization, as there is a high cost to pharmacomechanical interventions for the many U.S. patients develop DVT every year. Enter the ATTRACT trial.⁴ 

The ATTRACT trial is a National Institutes of Health–funded, multicenter, 1:1 randomized, open-label, assessor-blinded, controlled clinical trial. Patients with symptomatic proximal DVT were enrolled if symptoms had been present for less than 14 days. Half of the subjects underwent pharmacomechanical catheter–directed thrombolysis (intervention group) with both receiving initial and long-term anticoagulation (warfarin or a direct oral anticoagulant) consistent with published guidelines. Patients with cancer were not eligible. Pharmacomechanical thrombolysis consisted of recombinant tissue plasminogen activator (rtPA) up to a total dose of 35 mg administered into the thrombus by various delivery methods. This could be followed, if needed, by thrombectomy and optional balloon venoplasty and stent placement. Fifty-six U.S. centers participated. The degree of PTS and quality of life were assessed at various time points during two years of follow-up. PTS scales, in this case, the Villalta scale, score the severity of patient-reported symptoms (pain, cramps, heaviness, paresthesia, pruritus) and clinician-observed signs (pretibial edema, skin induration, hyperpigmentation, venous ectasia, redness, pain during calf compression). Scores range from zero to 33, with higher scores indicating more severe PTS.⁵ 

Primary efficacy outcome was the development of PTS of any severity or a leg ulcer at any time between six to 24 months follow-up. Key secondary efficacy outcomes were:1) the severity of PTS; 2) the proportion of patients with moderate to severe PTS; 3) changes in VCSS; and 4) vein health–related quality of life.

Study authors randomly assigned 692 patients and there was no difference in the primary outcome: PTS developed in 47 percent of patients assigned to pharmacomechanical thrombolysis and in 48 percent of patients in the nonintervention group (RR, 0.96; 95% CI, 0.82 to 1.11; p=0.56). However, major bleeding within 10 days of randomization occurred in 1.7 percent of the pharmacomechanical thrombolysis group and in 0.3 percent of the control group (p=0.049). With respect to the secondary outcomes, moderate to severe PTS occurred in 18 percent of patients with pharmacomechanical thrombolysis and in 24 percent of the control group (RR, 0.73; 95% CI, 0.54 to 0.98; p=0.035); the intervention decreased the severity of PTS at all visits between six and 24 months (Table 1); between baseline and 10 days, leg pain measured on a Likert scale was decreased (p=0.02), as did leg circumference measured in centimeters (p=0.02), comparing the intervention with the control group; there was no statistically significant difference in the change in quality of life between the two treatment groups. Finally, a prespecified subgroup analysis showed that PTS (Villalta scale score ≥ 5 or ulcer) occurred in 62 percent of the patients with iliofemoral DVT treated with intervention, in 65 percent of those in the control group, and in 51 percent and 63 percent, respectively, of patients with isolated femoropopliteal DVT, indicating no preferential benefit of the intervention in decreasing the overall occurrence of PTS in patients with iliofemoral DVT (RR, 0.89; 95% CI, 0.77-1.02; p=0.10).

Catheter-directed pharmacomechanical thrombolysis did not prevent the development of PTS in patients with acute proximal DVT, but increased the risk for major bleeding. This argues that pharmacomechanical thrombolysis should not be the primary go-to treatment for ALL patients with acute proximal DVT. The focus of the publication on the primary endpoint as the most meaningful leads to the sense that this study was a “negative study” as it failed to show benefit of pharmacomechanical thrombolysis.

There are few more noteworthy aspects of this study. First, the evaluation of secondary endpoints demonstrates that patients undergoing the intervention developed less moderate-to-severe PTS and a lesser degree of PTS (as measured by Villalta and VCSS scores) than patients in the control group. One would think this might be clinically relevant and desirable; interestingly, though, it did not translate to a difference in the quality of life assessment. While at first this seems surprising, some parameters of the Villalta scale score, particularly hyperpigmentation, venous ectasia, redness, and to some degree, edema (particularly if it is mild), are not likely to impact a patient’s wellbeing. Nevertheless, the number of patients needed to treat (NNT) with pharmacomechanical intervention to avoid one patient from having moderate-to-severe PTS (as assessed by the Villalta scale score with the limitations listed above), is 15.

Second, leg pain and swelling 10 to 30 days after enrollment were less likely in the intervention group than in the control group, indicating that the intervention leads to more rapid symptomatic improvement. This could translate to meaningful clinical benefit as it may lead to less time lost from work and other important daily activities, as well as decreased subsequent utilization of health-care resources such as follow-up visits for pain management. Third, the risk of major bleeding with the intervention was relatively low, and none of the major bleeding events were intracranial. The number needed to harm (NNH) a patient with the procedure by causing a major bleed is 70. Another key question is whether certain subgroups benefit from the intervention. This was not explored in this primary endpoint-focused publication. Nuances that might help clinicians with deciding who to consider for the intervention have, therefore, not yet been distilled out. Subsequent data analyses and publications will surely compare more in detail patients with iliofemoral DVT to patients with isolated femoropopliteal DVT, and patients with severe symptoms and higher Villalta and Venous Clinical Severity scores on presentation to patients with less severe symptoms and lower scores. It is, however, already foreseeable, that subanalyses will suffer from being underpowered. Nevertheless, they may well give hints as to which selected patients may benefit from pharmacomechanical intervention. Finally, given that many patients still developed PTS in spite of fairly successful thrombus removal, the study suggests that mechanisms other than the anatomical occlusion such as vascular inflammation, may contribute to the formation of PTS and may be worthwhile targets to explore in the future to decrease the development of PTS.