Risk Factors and Outcomes of Ischemic Stroke Associated With Tandem or Isolated Cervical Artery Occlusion in Anterior and Posterior Circulation

BACKGROUND: Tandem occlusion stroke, defined by concurrent extracranial and intracranial vessel occlusions, remains an underinvestigated subset of acute ischemic stroke. There is a critical gap in knowledge and data regarding tandem occlusion-related stroke treatment, mainly related to the very low inclusion rate of tandem stroke in endovascular thrombectomy trials, possibly in relation to the more severe features of tandem occlusion stroke. The HERMES meta-analysis only had 122 participants with tandem occlusions, preventing from any inference on treatment effect and leading to a critical gap in current management guidelines (ESO-ESMINT guideline on mechanical thrombectomy 2019; TUrc et al., 2019). While observational studies suggest potential benefits of emergent carotid stenting (eCAS), they also highlight risks such as hemorrhagic transformation and clinical deterioration, underscoring the need for large-scale data. Identification of features that could stratify the potential benefit of eCAS in patients with anterior circulation tandem occlusions could guide more individualised treatment strategies. CERES-TANDEM is a multicenter study designed to improve the understanding of tandem ischemic stroke, with particular regard to the impact of emergent carotid stenting, early antithrombotic treatment, and early bleeding. Objectives Identify stroke risk factors: The prevalence and adjusted associations of traditional vascular risk factors-hypertension, diabetes mellitus, hypercholesterolaemia, smoking status, atrial fibrillation, prior stroke or transient ischaemic attack (TIA), and pre-stroke modified Rankin Scale (mRS)-between patients presenting with tandem cervical-intracranial occlusions and those with isolated large-vessel occlusions will be compared. Multivariable logistic regression will estimate odds ratios (ORs) for tandem versus single-site occlusion, adjusting for relevant covariates (age, sex). Compare clinical outcomes of reperfusionIn both tandem-occlusion and isolated-occlusion cohorts, evaluating whether acute reperfusion therapies-intravenous thrombolysis, mechanical thrombectomy, and emergent carotid stenting-are associated with improved 3-month functional outcome. The estimate of the effect of each intervention versus medical management alone will be calculated using an inverse-probability-of-treatment-weighted (IPTW) ordinal logistic regression for shift in the full mRS distribution, and IPTW logistic regression for the dichotomous "good outcome" endpoint (mRS 0-2). Additional outcomes will adjust for baseline NIHSS, ASPECTS, onset-to-groin times, age and gender, and atrial fibrillation. Assess post-stent therapy. Among patients undergoing emergent carotid stenting, a comparison of three post-procedural antithrombotic regimens-single antiplatelet, dual antiplatelet, and anticoagulation-on functional outcome, symptomatic intracranial haemorrhage (ECASS II definition), and recurrent ischaemic events within 90 days. IPTW-weighted logistic regression for binary endpoints will be used, with regimen as the exposure and adjustment for stent type, procedural variables, and baseline risk factors. Study Design: A multicentre, observational cohort study pooling data from prospective stroke registries. Inclusion criteria: adults (≥18 years) with acute ischaemic stroke due to anterior-circulation large-vessel occlusion treated between January 2018 and December 2024. Tandem occlusion defined as simultaneous extracranial internal carotid and intracranial large-vessel occlusion on imaging. Ethics approval was obtained at each site if needed; all data were de-identified under encrypted study IDs. Data Collection: a predefined form available on solid support for offline use (constrined field dataset) and a RedCap CRFs will be available for data entry: * Baseline variables (e.g. demographics, vascular risk factors, pre-stroke mRS, NIHSS on admission, onset-to-imaging time, ASPECTS, occlusion site) * Imaging data: CT/CT-perfusion parameters, angiographic findings (TICI scores) * Procedural details: stent subtype, access route, thrombolysis use, thrombectomy device, stenting sequence (anterograde vs retrograde), procedural duration. * Medications * Outcomes of interest at follow-up visits. Data completeness and consistency were monitored centrally; missingness patterns will be evaluated and handled via multiple imputation (Rubin's rules for pooling). Statistical analysis: All analyses will employ IPTW to emulate a randomised comparison: propensity scores for each treatment comparison will be estimated by logistic regression incorporating prespecified confounders (demographics, baseline stroke severity, imaging features, centre, calendar time). Loveplots will be used to display balance. Ordinal outcomes: weighted proportional-odds models yielding common odds ratios (cORs) with 95% CIs; model fit and proportionality tested via score tests. Binary outcomes: weighted logistic regression producing ORs. Subgroup and interaction analyses using IPTW model will be conducted for: pre-specified for large/medium vessel occlusion, access route, intravenous thrombolysis, antithrombotic intensity, and sedation. Estimands: All analyses will be conducted in accordance with the estimand framework (ICH E9(R1)). The investigators define a total-effect estimand using treatment-policy strategy, a controlled direct-effect estimand adjusting for post-treatment recanalization and bleeding, and a principal-stratum estimand restricting the population to never-crossers. The primary outcome will be compared between eCAS strategy and a no-stenting strategy in patients with tandem-occlusion stroke. The primary estimand (treatment-policy centred) aims at identifying the effect of eCAS on mRS regardless of other interventional outcomes (total effect). The population includes all participants, with the eCAS group represented by those undergoing eCAS and no-stenting group including all those with any conservative treatment, regardless of any later crossover from the latter to the former. Stabilized weights will be applied in a weighted ordinal regression for the primary outcome to estimate the total effect (weighted ordinal regression; estimand 1), with results reported as common odds ratio (cOR) with 95% confidence intervals (CI) for shift in lower mRS score with eCAS vs no-stenting group. Secondary outcomes included excellent and good functional outcome. Additional outcomes (sICH, successful recanalization, END) were also calculated. The investigators predefined an interaction analysis for access route, approach, large/medium vessel occlusion, type of sedation, IVT and periprocedural antithrombotic treatment intensity in the context of estimand 1, fitting a cumulative proportional regression and adjusting for age, sex, pre-stroke mRS, NIHSS at admission, ASPECTS, atrial fibrillation and OTG. For estimand 1, the investigators prespecify a sensitivity analysis to account for the potential impact of successful intracranial recanalization on eCAS indication (reverse temporal association; sensitivity analysis on estimand 1). A second estimand to identify the effect of eCAS on mRS not mediated by successful reperfusion or sICH (direct effect, estimand 2) will also be implemented. In the full cohort, successful recanalization (TICI 2b-3) and sICH were included as factors in the IPTW-weighted ordinal regression, reporting the adjusted cORs for eCAS (estimand 2). A third principal-stratum estimand was planned, restricting the no-stenting group to participants who did not receive any treatment on the extracranial segment (excluding cases undergoing eCAS or any angioplasty procedure at the cervical occlusion site at the time of EVT or within 24 hours from EVT (estimand 3). THe investigators will use R v4.3 for all analysis.