ABSTRACT NUMBER: ESOC2026A2213 INCREASED RISK OF REPERFUSION INJURY BEYOND THE RANGE OF NORMOGLYCEMIA WITH A U-SHAPED RELATIONSHIP IN ENDOVASCULAR THOMBECTOMY FOR ACUTE ISCHEMIC STROKE

Abstract Background and aims Hyperglycemia has been suggested as a risk factor for haemorrhagic reperfusion injury after endovascular thrombectomy (EVT) in large vessel occlusion (LVO) stroke. However, the specific glucose thresholds impacting clinical outcomes remain unclear. We investigated the dose-response relationship between admission glucose and intracranial hemorrhage (ICH) in a large, real-world cohort. Methods We analysed data from our prospective thrombectomy registry of LVO patients treated from 01/2017 to 01/01/2023 at a German tertiary stroke centre. We evaluated the association between blood glucose levels at admission and risk of ICH as classified by the Heidelberg Bleeding Classification. Blood glucose levels were categorized in quintiles to account for nonlinearity. Causal effects were estimated applying inverse probability weighted regression analysis with adjustment for relevant demographic, clinical, laboratory (including hemoglobin A1c), and imaging characteristics. Results In our cohort of 1080 LVO patients (544 female [50.4%], median age 77 years [67-84, IQR]), haemorrhagic reperfusion injury occurred in 423 patients (39.2%). We identified a U-shaped relationship between admission glucose and ICH risk (Figure). Compared to an optimal blood glucose range (5.96-6.73mmol/L, ICH rate 33.3%), both low glucose (3.96-5.95mmol/L, average treatment effect (ATE) 14.5%, 95%CI[3,1%-25,8%], p=0.01) and high glucose levels (9.10-23.18mmol/L, ATE 13.1%, 95%CI[1,8%-24,5%], p=0.02) were independently associated with increased ICH risk. Conclusions Strict maintenance of normoglycemia prior to EVT—avoiding both hypo- and hyperglycemia—represents a promising strategy to reduce reperfusion injury; notably, this effect is independent of long-term glycemic control, underscoring the importance of glucose management during the hyperacute phase of stroke. Conflict of interest Annahita Sedghi: nothing to disclose, Norma Diel: nothing to disclose, Martin Arndt, Isabella Püschel: nothing to disclose, Kristian Barlinn: nothing to disclose, Daniel Kaiser: nothing to disclose, Timo Siepmann: nothing to disclose Figure 1 - belongs to Conclusions