Mechanical thrombectomy Cost-effectiveness of mechanical thrombectomy compared with standard treatment in patients with acute ischaemic stroke Robert Heggie1, Olivia Wu1, Keith Muir1, Phil White2, Gary Ford3, Martin M Brown4, Andrew Clifton5, Joanna Wardlaw6 1University of Glasgow, 2Newcastle University, 3Oxford University Hospitals NHS Trust, 4University College London, 5University of London, 6University of Edinburgh Background Eight recently published clinical trials1-8 have supported the use of mechanical thrombectomy (MT), using stent retrievers, in the endovascular treatment of acute ischaemic stroke. All the trials, with the exception of one1, were terminated early due to demonstrated efficacy. In 2014, the European Stroke Organisation (ESO)-Karolinska Stroke Update conference released a consensus statement, supporting the use of MT. Similarly, in 2015, the National Institute of Health and Care Excellence (NICE) guideline made a recommendation for the use of MT. “Mechanical thrombectomy, in addition to intravenous thrombolysis within 4.5 h when eligible, is recommended to treat acute stroke patients with large artery occlusions in the anterior circulation up to 6 h after symptom onset.” “Current evidence on the safety and efficacy of mechanical clot retrieval for treating acute ischaemic stroke is adequate to support the use of this procedure provided that standard arrangements are in place for clinical governance, consent and audit.” Aim To estimate the cost-effectiveness of mechanical thrombectomy in combination with IV-tPA compared with standard treatment (IV-tPA alone), in patients with acute ischaemic stroke. Methods A cost-effectiveness analysis of MT compared with IV-tPA, from the perspective of the UK NHS. The analysis was carried out over two time horizons: (i) 90 days based on the PISTE trial and (ii) lifetime, using all available evidence. 90-day within-trial analysis The Pragmatic Ischaemic Stroke Thrombectomy Evaluation (PISTE) trial was the only trial conducted in a UK setting. The primary outcome of the trial was the modified Rankin Scale (mRS). A mapping algorithm9 was used to convert mRS to health utilities, in order to calculate quality-adjusted life-years (QALYs). Unit costs were applied to resource use data collected within the trial. Generalised linear models were used to estimate mean costs and QALYs for each arm of the trial. Cost-effectiveness was expressed as incremental cost-effectiveness ratio (ICER). A non-parametric bootstrap was used to estimate the standard errors and 95% confidence intervals for the mean costs and QALYs. Lifetime economic model A model consisting of a decision tree (representing pathway for the first 3 months) and a 4-state Markov model10 (representing pathway beyond 3 months) were used. The probabilities mRS outcomes in each arm were estimated by meta-analysis of data from the 7 published trials of MT (data from THERAPY trial not available at time of publication). Transition probabilities between health states in the Markov model were sourced from other literature sources (data not shown). Cost of treatment and lifetime care were obtained from the literature. The costs of MT and IV-tPA alone were estimated to be £8111 and £1919, respectively. Model parameters Methods (continued) Sensitivity analysis was carried out to explore the impact of reperfusion within 3 hours and 8 hours. Our basecase included patients treated up to 12 hours following stroke onset. Probabilistic sensitivity analysis was carried out. Value of implementation To estimate the value of implementation, we compared the net monetary benefit (potential health benefit in monetary terms) of MT compared with the cost of implementing this procedure into routine practice within the UK. We assumed a technology life of 5 years. Based on an estimation of 9,100 eligible patients per year (discounted at 3.5%), we estimated that 42,525 patients who could potentially benefit over this period. Results Mechanical thrombectomy (MT) in combination with IV-tPA was associated with higher costs, but greater health benefits, when compared with IV-tPA. Over a 90-day period, MT was associated with an additional cost of £5,207 and an additional QALY gain of 0.0154. The ICER was £338,117. Similarly, over the horizon of a lifetime, MT was associated with greater costs and health benefits than IV-tPA. However, the ICER was substantially reduced - £3,857 (see Table). Sensitivity analysis showed that ICERs increase with increasing time to reperfusion. The net monetary benefit (health benefit in monetary terms) is £13,704 per patient. Assuming a 5-year time horizon and full implementation, the value of implementation was £433 million. If implementation is greater than 26%, the value of implementation is greater than the cost of implementation. Parameters Values Decision tree: mRS 0-2 (MT) 0.57 Health utility: Independent 0.74 Decision tree: mRS 3-5 (MT) 0.27 Health utility: Dependent 0.38 Decision tree: mRS 6 (MT) 0.16 Health utility: Recurrent 0.34 Decision tree: mRS 0-2 (IV-tPA) 0.26 Cost first 3 months: Independent £7,303 Decision tree: mRS 3-5 (IV-tPA) 0.55 Cost first 3 months: Dependent £15,627 Decision tree: mRS 6 (IV-tPA) 0.19 Cost first 3 months: Death £10,039 Model Structure Results of lifetime model Mechanical thrombectomy IV-tPA Inc Costs Inc QALYs ICERs Cost QALY gained QALY gained Base case £40,131 5.36 £34,561 3.92 £5,569 1.44 £3,857 Reperfusion within 3 hours £38,212 5.56 £4,653 1.71 £2,682 Reperfusion within 8 hours £41,279 4.81 £6,717 0.96 £6,914 Conclusion Mechanical thrombectomy (MT) in combination with IV-tPA was associated with higher costs, but greater health benefits, when compared with IV-tPA alone. Based on the time horizon of 90-days, MT was not cost-effective compared with IV-tPA. However, this result was reversed if we consider a lifetime horizon. The long-term health benefits associated with MT outweigh the additional costs, compared with IV-tPA. In line with clinical evidence, cost-effectiveness improves further as time-to-treatment reduces. We estimate that the value of implementation is greater than the cost of implementation of MT into routine practice is the UK. The results holds if we assume full implementation or 50% implementation. Mechanical thrombectomy has recently been approved by NHS England and it is estimated that potentially 9,100 patients per year could benefit from this treatment. Contact details: Robert Heggie , Email: robert.heggie@glasgow.ac.uk 1. Berkhemer et al, NEJM 2015; 372:11-20. 2. Goyal et al, NEJM 2015; 372:1019-30. 3. Campbell et al, NEJM 2015; 372:1009-1018. 4 .Saver et al, NEJM 2015; 372:2285-2295. 5. Jovin et al, NEJM 2015; 372:2296-2306. 6. Mocco, Int J Stroke 2015; 10:10. 7. Bracard et al, Lancet Neurology 2016; 11:138–1147. 8. Muir et al, J Neurol Neurosurg Psychiatry 2017; 88:38-44. 9. Rivero-Arias et al, Med Decis Making 2010; 30:341-54. 10. Ganesalingam et al, Stroke 2015; 46:2591-8