Cost Effectiveness of Early Endovenous Thrombus Removal for Acute Iliofemoral Deep Vein Thrombosis in the United Kingdom.

OBJECTIVE: Early clot removal using endovascular intervention aims to reduce post-thrombotic syndrome (PTS) following iliofemoral deep venous thrombosis (DVT). This may reduce long term morbidity but incurs a higher initial cost. This study examined the cost effectiveness of catheter-directed thrombolysis (CDT) and pharmacochemical thrombectomy (PMT) compared with oral anticoagulation (OAC) alone for treatment of acute iliofemoral DVT in the United Kingdom.

METHODS: A combined decision tree (acute DVT complications) and Markov model [long term complications (PTS)] was used decision analytic modelling with five states: no PTS, mild PTS, moderate PTS, severe PTS, and dead. All patients started with acute DVT. Patients who survived acute complications transitioned into the Markov model. Cycle time was 6 months. A healthcare payer perspective and lifetime horizon was used, adjusting for excess mortality due to history of thrombosis. Data for probabilities, transition probabilities, mortality, and utilities were obtained from published literature. Cost data were obtained from UK NHS tariffs and published literature. Outcomes were mean lifetime cost, quality-adjusted life years (QALYs), and cost effectiveness.

RESULTS: Over a patient's lifetime, OAC was more costly (£37 206) than CDT (£32 043) and PMT (£36 288). Mean lifetime QALYs for OAC (12.9) were lower than CDT (13.5) and PMT (13.3). Therefore, in the incremental cost effectiveness analysis, both CDT and PMT were dominant: CDT was less costly (- £5 163) and more effective (+0.6 QALYs) than OAC, and PMT was also less costly (-£917) and more effective (+0.3 QALYs) than OAC. Results were robust to univariate sensitivity analyses, but probabilistic sensitivity analyses suggested considerable parameter uncertainty.

CONCLUSION: Early interventional treatment of iliofemoral DVT is cost effective in the UK. Future clinical and epidemiological studies are needed to characterise parameter uncertainty. Further analysis of modern practice, alternative treatments, and optimised care models is warranted.