Analysing the results of a retrospective, single-centre cohort study conducted at Guys and St Thomas’ Hospital, London, UK, Anna Pouncey (Guys and St Thomas’ Hospital, London, UK) emphasised that over 50% of cases included in the investigation could have been avoided. When questioned about how the findings of this study will impact practice, Pouncey said: “We now have a lower threshold for investigating and checking if a patient is showing signs of restenosis.”
Presented at the European Society for Vascular Surgery annual meeting (ESVS 2019; 24–27 September, Hamburg, Germany), this study into risk factors for patients requiring a reintervention, following treatment for acute iliofemoral deep vein thrombosis (DVT), was based on a classification model that highlighted technical, flow-related and haematological causes of restenosis and re-occlusion.
Furthermore, Pouncey also offered a cost-effectiveness analysis of interventions for the treatment of iliofemoral DVT and subsequent post-thrombotic syndrome (PTS), suggesting that either pharmacomechanical thrombectomy (PMT) or catheter-directed thrombolysis (CDT) may present benefits over oral anticoagulation.
Classification of risk factors for reintervention
“The aim of our study was to develop a classification system for why patients require a reintervention, identify the individual risk factors associated with them, and examine patient outcomes following reintervention,” said Pouncey, as she outlined the objectives of St Thomas’ investigation.
As part of the study, 135 patients who had undergone successful lysis for acute symptomatic iliofemoral DVT—between 2013 and 2017—were identified, with 33% of these patients requiring a reintervention. In addition, patient notes were evaluated for risk factors (including the haematologist’s assessment of their anticoagulation status), while a retrospective venographic assessment was also performed with the support of two independent specialists.
Focusing on the data, Pouncey revealed that 54% of patients (who had undergone a reintervention) had a technical error identified, which could have been a lack of stenting in the presence of residual disease, a failure to extend the stent distally (missing inflow), a failure to extend the stent proximally (missing outflow), or device failure such as stent fracture. However, Pouncey noted that “very few technical issues were related to device failure”. Around 44% had a flow-related issue, which would entail scarring or an occlusion of the femoral, profunda, or popliteal vein, and 35% had a haematological issue, such as non-compliance with anticoagulation or inappropriate dosing.
Of all the cases included in the study, 29% were multifactorial, with a significant amount of overlap between technical and flow-related issues. Comparing this to the cohort who maintained their vessel patency, Pouncey commented: “We can see that very few of these patients had any technical issues, all had good flow, and only one had a haematological issue.”
Looking next at patient factors, or individual risk factors associated with reintervention, it was stated that the only relevant variable was a younger patient age, although these patients were associated with a greater prevalence of inferior vena cava (IVC) thrombosis. In terms of flow-related factors, profunda occlusion, femoral occlusion and stenting across the inguinal ligament were all considered important, while the patient was three times more likely to require an intervention if they were non-compliant with anticoagulation. “When we performed multivariate analysis, the three factors that came out as dominant were IVC thrombosis, stenting across the inguinal ligament and anticoagulation non-compliance,” Pouncey asserted.
Patient outcome data showed that the median time to reintervention was 45 days, and 45% of cases were reintervened on during the first six weeks. There was an intervention success rate of 67%, although this was much higher in cases where a reintervention was performed prior to total vessel occlusion. Regarding the incidence of PTS, it was found that the need for reintervention doubled the risk, with 32% of patients in the reintervention group developing PTS compared to only 14% in the cases that remained patent.
In summary, Pouncey said: “Anticoagulation compliance is essential, and there is a role for surveillance so that you can reintervene prior to occlusion and have a greater chance of success. Last of all, you have a greater risk of poor patient outcome if you have to reintervene, so try to get it right the first time.”
A cost-effectiveness analysis of interventions
“We all know that iliofemoral DVT and the subsequent incidence of PTS carries a significant financial burden,” commented Pouncey, “and if you perform an interventional treatment to expedite thrombus removal, you will hopefully reduce the patients’ symptoms and mitigate post-thrombotic syndrome.” However, the investigator also accepted that this comes with “a higher risk of bleeding” as well as a greater cost.
Moreover, there is only a modest health benefit to this intervention (a 10% reduction in moderate to severe PTS) according to the ATTRACT trial. Despite these findings, PMT is currently used in the UK, albeit under special measures due to there being “a paucity of data to support its use”, according to Pouncey. “We aimed for an NHS-based cost-effectiveness analysis for treatment of iliofemoral DVT and specifically wanted to look at all anticoagulation vs. percutaneous interventions,” she said.
In order to perform the cost-effectiveness analysis, a patient health states model was created; a front-end decision tree was used to model the acute DVT veins and any complications that could happen, while a backend Markov model was also used to model the long-term patient outcomes, namely the severity of PTS. Analysis was conducted through a lifetime horizon with death as the absorbing state, while slightly increased mortality rates following DVT were also accounted for.
Pouncey addressed the sourcing of data and explained that “for patient outcomes, we used the results of the ATTRACT trial, as well as for the proportions of patients in each health state over time”. She added: “For our utility estimates we used the best available published literature and for our costing sources, we used the UK NHS national tariff workbook, published scientific literature and local data from our team at the St Thomas’ Trust.”
The results of the analysis, which accounted for a stenting rate of 80%, were in some ways unsurprising, demonstrating that oral anticoagulation is far cheaper than CDT or PMT. However, the main results of the model established as part of the investigation at St Thomas’ Hospital showed that both CDT and PMT were dominant over oral anticoagulation. Concerning the findings, Pouncey commented: “What this means is that over a lifetime, the mean cost for the interventional treatments was cheaper than oral anticoagulation and the patient quality of life was higher too. So we looked into this a bit further, and what we found was that the increased cost is offset by a long-term cost benefit in the reduction of moderate to severe PTS. Moderate and severe PTS also have much higher healthcare costs over the long-term.”
In response to these results, the robustness of the model was assessed univariate and multivariate sensitive analyses. This revealed that costing is most sensitive to age, stenting and hospitalisation for severe PTS and that clinical effectiveness is more sensitive to major bleeding rates than pulmonary embolism (PE). That being said, significant parameter uncertainty was also found, reflecting the quality of data which was put into the cost-effectiveness model.
Pouncey concluded: “CDT and PMT may be cost-effective over oral anticoagulation, due to a reduction in the severity of PTS, however there is still uncertainty and we need to address this before drawing further conclusions. This can be achieved with more robust data on intervention outcomes, as well as a better understanding of the cost and lifetime impact of severe PTS.”