{
  "abstract": "Background Racial disparities in endovascular thrombectomy (EVT) for acute ischemic stroke are well documented, with Black patients receiving EVT less frequently than White patients. Prior analyses have adjusted for stroke severity as a potential confounder but have not examined whether this disparity varies across the severity spectrum, where treatment decisions carry the greatest consequence. We aimed to investigate whether racial disparities in EVT use change with stroke severity and to characterize the interaction between race, sex, and stroke severity on treatment access.Methods We conducted a cross-sectional study using the National Inpatient Sample (NIS, 2018-2022), identifying adults with acute ischemic stroke (ICD-10-CM code I63) at US teaching hospitals. Acute ischemic stroke cases were stratified by race (Black vs White) and sex. Stroke severity was assessed using National Institutes of Health Stroke Scale (NIHSS) scores extracted from ICD-10-CM codes (R29.700-R29.742). EVT was identified using ICD-10-PCS codes for intracranial, carotid, and vertebral artery extirpation. Survey-weighted logistic regression with restricted cubic splines modeled EVT probability as a function of NIHSS score, including 3-way interactions between NIHSS, race, and sex. Models were adjusted for age, insurance status (Medicare, Medicaid, private, uninsured), comorbidities (hypertension, diabetes mellitus, atrial fibrillation, heart failure), intravenous thrombolysis, calendar year, and geographic region. Marginal effects were estimated at each NIHSS level. Likelihood ratio tests assessed interaction significance. Model discrimination was evaluated using the C-statistic.Results Among 325,799 patients (White men 128,684; White women 122,693; Black men 36,910; Black women 37,512), NIHSS was documented in 59.1-63.4% across groups. Black patients were younger (mean age 62.9-66.3 vs 69.4-73.7 years), had higher prevalence of hypertension (89.6-90.2% vs 84.1%), diabetes (45.6-51.1% vs 32.1-37.3%), and heart failure (20.8-22.3% vs 17.2-17.3%), but markedly lower atrial fibrillation prevalence (15.7-17.3% vs 27.7-30.5%) and lower rates of intravenous thrombolysis (11.3-11.4% vs 14.2-14.8%). Black patients received EVT less frequently overall (5.8-6.0% vs 7.1-7.8%). The race-sex-NIHSS interaction was significant (likelihood ratio test, p=0.002), indicating that disparities varied by stroke severity. Black patients had lower predicted EVT probabilities at every NIHSS level, and this gap widened with increasing severity: the disparity averaged 1 percentage point at NIHSS 5 (mild stroke), 3 percentage points at NIHSS 10 (moderate stroke), and 5 percentage points at NIHSS 20 (severe stroke). At NIHSS 20, predicted EVT probability was 33% for White men, 32% for White women, 28% for Black men, and 26% for Black women. Black women had the lowest predicted treatment probability at every severity level. The model demonstrated good discrimination (C-statistic 0.853, 95% CI 0.851-0.856).Conclusion Racial disparities in EVT widened progressively with stroke severity, with Black women having the lowest predicted treatment probability at every level. This severity-dependent gradient is consistent with fundamental cause theory, where larger disparities occur for conditions more amenable to effective treatment. Equity monitoring in stroke systems should be stratified by severity, as aggregate EVT rates may mask widening disparities among patients with the greatest potential to benefit.Disclosures M. Essibayi: None. H. Jamil: None. J. Hirsch: None.",
  "authors": [
    {
      "affiliations": [
        "Neurosurgery, Albert Einstein College of Medicine, Bronx, NY"
      ],
      "name": "M Essibayi"
    },
    {
      "affiliations": [
        "Neuroradiology, West Virginia University, Morgantown, WV"
      ],
      "name": "H Jamil"
    },
    {
      "affiliations": [
        "Massachusetts General Hospital, Boston, MA"
      ],
      "name": "J Hirsch"
    },
    {
      "affiliations": [
        "Neuroradiology, West Virginia University, Morgantown, WV"
      ],
      "name": "D Lakhani"
    }
  ],
  "title": "O-043 Racial disparities in endovascular thrombectomy widen with stroke severity: a national inpatient sample analysis",
  "uid": "9f98900e-baee-54c5-8b59-0110e9f92fbe"
}