We propose no update of the NICE guideline on venous thromboembolic diseases: diagnosis, management and thrombophilia testing (CG144).
The purpose of this exceptional review was to examine any impact on NICE's guideline on venous thromboembolic diseases following the publication of the pharmacomechanical catheter-directed thrombolysis for deep-vein thrombosis (ATTRACT) study. This study was considered to be relevant to the section on thrombolytic therapy treatment for deep vein thrombosis in NICE guideline CG144. Additional evidence in this area from previous surveillance published since the publication of NICE's guideline on venous thromboembolic diseases in June 2012 was also considered by the exceptional review.
The ATTRACT study is a multicentre randomised controlled trial (56 US clinical centres) that compared the effectiveness of anticoagulation plus pharmacomechanical catheter-directed thrombolysis (PCDT) to anticoagulation alone (control group). Patients aged between 16 and 75 years were included if they had symptomatic proximal deep vein thrombosis (DVT) in the femoral, common femoral or iliac vein. People were excluded if their symptoms lasted more than 14 days, were pregnant, had active cancer, had DVT or post-thrombotic syndrome (PTS) in the past 2 years, had active bleeding or were at an increased risk of bleeding, or a life expectancy of less than 2 years.
Each person was randomly allocated to 1 of the 2 interventions by a web-based system to conceal allocation. All patients received anticoagulant therapy and were provided with elastic compression stockings before randomisation. Those patients in the intervention group received PCDT in a 5‑phase procedure, which allowed variation in the delivery method of the mechanical aspect of catheter-directed thrombolysis dependent on the location and extent of the thrombus. All those in the PCDT group received initial delivery of alteplase – a recombinant tissue plasminogen activator (rt‑PA) – under strict dosing parameters as described in the trial protocol. The need and type of further procedures required to remove any residual thrombus was left to the discretion of the clinicians.
The main efficacy outcome assessed was development of PTS measured using the Villalta scale. A score of 5 or greater on the Villalta scale indicated the presence of PTS and scores were categorised to provide severity ratings of mild, moderate and severe. The secondary efficacy outcome assessed was severity of PTS measured using the Venous Clinical Severity Score (VCSS). The VCSS was scored on 8 clinical items and 1 pain symptom item to give a total score ranging from 0 to 27 with a higher score indicating increased severity. For both scales, clinical examiners administering the measures were blinded to treatment allocation and measures were taken at 6, 12, 18 and 24 months.
The other secondary outcomes assessed were: health-related quality of life at baseline and 24 months using the VEINES instrument and the Short Form‑36 Health Survey; a major non‑PTS treatment failure defined as having either surgical intervention for venous disease within 6 months, amputation of the leg within 24 months, or development of venous gangrene within 6 months; and presenting DVT symptoms of leg pain and swelling at baseline, 10 days and 30 days. Venograms of patients in the PCDT intervention group were assessed before and after the procedure for the degree to which the thrombus had been removed. The study also measured safety outcomes including bleeding, symptomatic pulmonary embolism, symptomatic recurrent DVT and death.
All outcomes were analysed using a modified intention-to-treat analysis and a per-protocol analysis along with a sensitivity analysis to account for missing follow‑up assessments.
A total of 692 people were randomised (337 in the PCDT group, 355 in the control group). One person from the PCDT group was subsequently excluded from analysis because they did not have a qualifying DVT. No differences were identified in the baseline characteristics between the groups.
In the intention-to-treat analysis, no differences in the development of PTS over the 24‑month period between the PCDT group (157 out of 336 developed PTS) and the control group (171/355) were identified (risk ratio 0.96, 95% confidence interval [CI] 0.82 to 1.11, p=0.56).
The per-protocol analysis excluded 16 people (5 assigned to the control group and 11 assigned to the PCDT group) who did not undergo the intervention assigned at randomisation. The analysis found no differences in the development of PTS over the 24‑month period between the PCDT group (151/325) and the control group (169/350; risk ratio 0.94, 95% CI 0.81 to 1.10, p=0.47). Similar results were found in the sensitivity analysis (risk ratio 0.89, 95% CI 0.78 to 1.02).
Both intention-to-treat and per-protocol analyses found significantly lower severity of PTS in the PCDT group compared with the control group at the 6-, 12-, 18- and 24‑month follow‑up points for both the Villalta and VCSS scales (p<0.01). However, 1 exception was the scores on the VCSS scale at the 24‑month period, which was not significantly different between groups.
Neither analyses found significant differences between groups for any of the remaining secondary outcomes of health-related quality of life, major non‑PTS treatment failure, overall treatment failure or symptoms of DVT (all p>0.01).
Analysis of safety outcomes indicated increased rates of major bleeding in the PCDT group (6/336) compared to the control group (1/355) within the first 10 days (risk ratio 6.18, 95% CI 0.78 to 49.2, p=0.049). However, the difference was no longer significant over the total 24‑month period with 19 instances of major bleeding in the PCDT group compared to 13 instances in the control group (risk ratio 1.52, 95% CI 0.76 to 3.01, p=0.23).
A significant difference was found for rates of any bleeding (non-major) within the first 10 days with 15 instances in the PCDT group compared to 6 instances in the control group (risk ratio 2.64, 95% CI 1.04 to 6.68, p=0.03). Over the total 24‑month period, the difference was no longer significant with 46 instances in the PCDT group compared to 38 instances in the control group (risk ratio 1.26, 95% CI 0.85 to 1.89, p=0.25).
No significant differences were found between groups at 10 days or over 24 months for the other safety outcomes of recurrent DVT or deaths. Although no deaths occurred in either group within the first 10 days, 1 incidence of death due to pulmonary embolism was recorded in the PCDT group at 6 months.
During guideline development, it was recognised that clinical practice in the treatment of DVT had moved towards using catheter-directed administration of thrombolytic agents for clot lysis rather than systemic administration. However, it was also noted that the evidence was unclear whether catheter-directed thrombolytic therapy reduces the incidence of PTS but that there may be an increased risk of major bleeding.
In assessing the evidence, the guideline considered mortality, risk of bleeding, recurrence of venous thromboembolism (VTE), and PTS as patient important outcomes. One Cochrane review that included 12 randomised controlled trials (RCTs) and 2 additional RCTs were evaluated relevant to this intervention and outcomes. The evidence showed that, compared to standard anticoagulation, catheter-directed therapy reduced all‑cause mortality (4 studies, relative risk [RR] 0.46, 95% CI 0.13 to 1.57), reduced recurrent VTE (3 studies, RR 0.19, 95% CI 0.05 to 0.7), reduced PTS (2 studies, RR 0.81, 95% CI 0.6 to 1.11), and increased rates of major bleeding (6 studies, RR 1.28, 95% CI 0.52 to 3.12).
However, the quality of evidence was considered low to moderate due to the high risk of bias of the studies included and the imprecision of the results found. The results for all‑cause mortality and major bleeding were deemed to be very uncertain whether they indicate clinically important differences between interventions. The 2 studies evaluating PTS showed that it is unlikely that there is any difference of clinical importance. Although it was only the 3 studies showing fewer instances of recurrent VTE that were considered to demonstrate a clinically important difference.
No evidence was identified that compared mechanical thrombectomy with standard anticoagulation or traditional thrombolysis.
A cost-effectiveness analysis was not conducted but the costs and the consequences of the interventions were considered, leading to the decision that the additional costs incurred by the use of catheter-directed therapy were lower than the cost associated with treating PTS.
An evidence update (2014) found 1 RCT and 1 Cochrane systematic review relevant to this intervention. The randomised controlled CaVenT study compared catheter-directed thrombolytic therapy plus standard anticoagulation with standard anticoagulation therapy alone in 209 people with DVT above mid-thigh level. At 24 months, significantly fewer people had PTS in the catheter-directed therapy group (p=0.047). However, catheter-directed therapy was also associated with bleeding complications. The Cochrane systematic review of 17 RCTs compared any type of thrombolytic therapy (systemic, loco-regional or catheter-directed) in addition to anticoagulation therapy with anticoagulation therapy alone in people with DVT of the lower limb. Thrombolytic therapy was associated with a significantly lower incidence of PTS (between 1 and 6 years, p=0.000026; 4 studies, n=341) and increased incidences of bleeding complications (p=0.00064; 17 studies, n=1,103).
The previous surveillance review (2016) found 3 studies relevant to this intervention. An RCT of people with iliofemoral DVT found a significant reduction in risk of deep venous reflux and an increased venous patency with the addition of catheter-directed therapy to anticoagulation and compression stockings at 24‑month follow‑up. The authors concluded that these outcomes were found to be strong predictors of PTS. A meta-analysis of 6 studies in people with proximal DVT found significantly lower rates of PTS and increases in major bleeding episodes with the addition of catheter-directed therapy compared with anticoagulation alone. However, catheter-directed therapy was not found to reduce mortality, recurrent DVT or PE. A 5‑year follow‑up of the CaVenT study compared standard treatment with compression stockings and anticoagulants (control group) to standard treatment plus catheter-directed therapy with alteplase in people (n=176) with proximal DVT. At 5‑year follow‑up, a significant reduction in risk of PTS was found for the catheter-directed therapy group.
We engaged with topic experts who helped to develop the guideline and those involved in the current update of NICE guideline CG144. Two topic experts commented that it is not possible to generalise the trial findings specifically to the relevant recommendation in the guideline because the ATTRACT study did not report subgroup analysis of the iliofemoral DVT population. Topic experts also commented that a significant proportion of participants had a long duration of rt‑PA infusion and that some patients were offered direct oral anticoagulants whereas others were not. These comments support the exceptional review proposal that the population and intervention within the ATTRACT study are not fully relevant to those covered in recommendation 1.2.6 in NICE guideline CG144. One topic expert also highlighted that the 5‑year follow‑up data to the CaVenT study are relevant to this recommendation and should be considered in the review. This study has been included in the review of evidence for this exceptional review.
Recommendation 1.2.6 currently states that catheter-directed thrombolytic therapy should be considered for patients with symptomatic iliofemoral DVT. To be considered for the intervention, these patients should have symptoms of less than 14 days' duration, good functional status, a life expectancy of 1 year or more, and a low risk of bleeding.
The new evidence from the ATTRACT study showed that PCDT did not prevent PTS in people diagnosed with acute proximal DVT but there was some evidence of more major bleeds within the first 10 days in this group compared to the anticoagulation alone control group.
Regarding the strengths and limitations of the ATTRACT study, the use of randomisation, allocation concealment and assessor blinding reduce potential selection and detection bias. The study included relevant patient important outcomes, all of them relevant for the guideline and considered in the development of the recommendation.
The population is partly relevant to the guideline with the inclusion of people with iliofemoral DVT who meet the criteria set out in the guideline recommendation. However, the study also includes people with femoropopliteal DVT and it was not possible to analyse subgroups because the study was not adequately powered to do so. The intervention used in the study combines mechanical thrombectomy with catheter-directed thrombolysis. Because this is a variation of the broader intervention considered in the recommendation, it may not be possible to generalise the results to all types of catheter-directed thrombolytic therapy. The authors also note that a limitation of the study was that the mechanical aspect of the procedure varied according to clinical need and clinician preferences. The degree to which this limitation impacted on the results of the main and safety outcomes remains unclear because randomisation was not performed for this aspect of the procedure to allow separate analysis. Further variations in the intervention included the duration of rt‑PA infusion and that some patients received direct oral anticoagulants whereas others did not. Again, it is unclear what role these variations played in the overall effectiveness of the intervention.
The loss of follow‑up was more than the anticipated 10% and primarily occurred in the control group, which the authors recognise may result in an underestimate of the treatment effect. However, statistical analyses were conducted appropriately to account for missing data and crossover between interventions which found similar results.
The result which showed an increase in bleeding associated with PCDT highlighted a risk already identified in previous studies for this population. However, there is some inconsistency between the confidence intervals and p value reported for this result. Although the p value may indicate significance, albeit borderline, the confidence intervals are wide and includes the null value. Also, although it appears that bleeding risk increased in the first 10 days, at 24 months this difference was no longer significant between groups and no deaths occurred in the initial 10‑day period. So although there is some suggestion of an increase in bleeding, the number of outcomes are small and the statistical results inconsistent.
The recommendation on catheter-directed thrombolytic therapy in NICE guideline CG144 was developed with some prerequisite criteria due to the unclear evidence base. Evidence relevant to the population and intervention published since the development of the guideline generally supports this advice but should be approached with some caution due to the imprecision in the results. Although the new evidence from the ATTRACT study does not show any benefit of PCDT in reducing PTS, it is unlikely to impact on NICE guideline CG144 due to the limitations in population and intervention identified.
This exceptional surveillance review did not search for new evidence relating to other clinical areas in the guideline.
See how we made the decision for further information.
This page was last updated: 30 August 2018