Reduced Impact of Endovascular Thrombectomy on Disability in Real-World Practice, Relative to Randomized Controlled Trial Evidence in Australia

Background and Aims: Disability-adjusted life years (DALYs) are an important measure of the global burden of disease that informs patient outcomes and policy decision-making. Our study aimed to compare the DALYs saved by endovascular thrombectomy (EVT) in the Australasian-based EXTEND-IA trial vs. clinical registry data from EVT in Australian routine clinical practice.

Methods: The 3-month modified Rankin scale (mRS) outcome and treatment status of consecutively enrolled Australian patients with large vessel occlusion (LVO) stroke were taken from the International Stroke Perfusion Imaging Registry (INSPIRE). DALYs were calculated as the summation of years of life lost (YLL) due to premature death and years lived with a disability (YLD). A generalized linear model (GLM) with gamma family and log link was used to compare the difference in DALYs for patients receiving/not receiving EVT while controlling for key covariates. Ordered logit regression model was utilized to compare the difference in functional outcome at 3 months between the treatment groups. Cox regression analysis was undertaken to compare the difference in survival over an 18-year time horizon. Estimated long-term DALYs saved based on the EXTEND-IA randomized controlled trial (RCT) results were used as the comparator.

Results: INSPIRE patients who received EVT treatment only achieved nominally better functional outcomes than the non-EVT group (p = 0.181) at 3 months. There was no significant survival gain from EVT over the first 3 months of stroke in both INSPIRE and EXTEND-IA patients. However, measured against no EVT in the long-term, EVT in INSPIRE was associated with no significant survival gain [hazard ratio (HR): 0.92, 95% confidence interval (CI): 0.78–1.08, p = 0.287] compared with the survival benefit extrapolated from the EXTEND-IA trial (HR: 0.42, 95% CI: 0.22–0.82, p = 0.01]. Offering EVT to patients with LVO stroke was also associated with fewer DALYs lost (11.04, 95% CI: 10.45–11.62) than those not receiving EVT in INSPIRE (12.13, 95% CI: 11.75–12.51), a reduction of −1.09 DALY (95% CI: −1.76 to −0.43, p = 0.002). The absolute magnitude of the treatment effect was lower than that seen in EXTEND-IA (−2.72 DALY reduction in EVT vs non-EVT patients).

Conclusions: EVT for the treatment of LVO in a registry of routine care was associated with significantly lower DALYs lost than medical care alone, but the saved DALYs are less than those reported in clinical trials, as there were major differences in the baseline characteristics of the patients.