{
  "abstract": "Background Early neurological improvement following reperfusion therapy may provide valuable prognostic information after acute large vessel occlusion stroke (LVOS). However, the relative predictive value of different National Institutes of Health Stroke Scale (NIHSS)-derived metrics for functional outcomes and mortality remains unclear.Methods We conducted a retrospective analysis of a prospectively maintained stroke database, including patients with anterior circulation LVOS who received bridging intravenous thrombolysis within 4.5 hours of last known well, with or without subsequent endovascular therapy, between January 2020 and August 2024. Baseline clinical, imaging, and workflow characteristics were evaluated. NIHSS-based metrics—including baseline NIHSS, 24-hour NIHSS, absolute NIHSS improvement, and percent NIHSS improvement—were assessed for their association with 90-day good functional outcomes (modified Rankin Scale [mRS] 0-2) and 90-day mortality (mRS 6). Absolute NIHSS improvement was defined as the difference between baseline and 24 hour NIHSS scores, while percent NIHSS improvement was calculated as the relative reduction in NIHSS from baseline to 24 hours, expressed as a percentage. Discrimination for predicting outcomes was assessed using C-statistics and DeLong’s test. Calibration performance was summarized by the calibration intercept, slope, and brier score. Multivariable logistic regression models adjusted for demographic, clinical, and treatment-related covariates were used to evaluate independent associations.Results Among 299 patients, 168 (56.2%) achieved 90-day good functional outcomes. Patients with poor outcomes were older (74 vs 69 years, p<0.001), had higher admission glucose (127 vs 121.5 mg/dL, p=0.009), higher baseline NIHSS (19 vs 14, p<0.001), and lower ASPECTS (9 vs 10, p=0.002). Favorable collateral status was strongly associated with good outcomes (85.2% vs 61.4%, p<0.001). In adjusted models, 24-hour NIHSS demonstrated the strongest discrimination for predicting good functional outcome (AUC 0.900, 95% CI 0.836-0.936), followed by percent improvement (AUC 0.874, 95% CI 0.820-0.914), absolute improvement (AUC 0.816, 95% CI 0.758-0.865), and baseline NIHSS (AUC 0.795, 95% CI 0.739-0.844). Similar patterns were observed for mortality prediction, where 24-hour NIHSS showed the highest discrimination (AUC 0.866, 95% CI 0.786-0.925), followed by percent improvement (AUC 0.826, 95% CI 0.749-0.888), absolute improvement (AUC 0.797, 95% CI 0.720-0.865), and baseline NIHSS (AUC 0.787, 95% CI 0.703-0.852). In multivariable models, lower baseline and 24-hour NIHSS and greater absolute and percent NIHSS improvement were independently associated with improved functional outcomes and lower mortality. All metrics demonstrated good calibration, with the 24-hour NIHSS showing the best performance (90-day mRS 0-2: Brier score = 0.133 and 90-day mRS 6: Brier score = 0.112).Conclusions NIHSS-derived metrics demonstrate strong prognostic performance following bridging thrombolysis in LVOS. This is unsurprising, given that the NIHSS captures early neurological response and infarct burden, both of which are closely linked to clinical outcomes. Among these metrics, the 24-hour NIHSS provides the highest predictive discrimination and calibration for both 90-day good functional outcomes and mortality, supporting its role as a pragmatic post-reperfusion assessment tool.Disclosures K. Darko: None. M. Cao: None. G. James: None. J. Kocherzat: None. S. Kozloski: None. C. Neal: None. C. Schirmer: None.",
  "authors": [
    {
      "affiliations": [
        "Geisinger Medical Center, Danville, PA"
      ],
      "name": "K Darko"
    },
    {
      "affiliations": [
        "School of Medicine, Geisinger College of Health Sciences, Scranton, PA"
      ],
      "name": "KE Kline"
    },
    {
      "affiliations": [
        "Geisinger Commonwealth School of Medicine, Scranton, PA"
      ],
      "name": "M Cao"
    },
    {
      "affiliations": [
        "Geisinger Commonwealth School of Medicine, Scranton, PA"
      ],
      "name": "G James"
    },
    {
      "affiliations": [
        "Geisinger Commonwealth School of Medicine, Scranton, PA"
      ],
      "name": "J Kocherzat"
    },
    {
      "affiliations": [
        "Geisinger Commonwealth School of Medicine, Scranton, PA"
      ],
      "name": "S Kozloski"
    },
    {
      "affiliations": [
        "Geisinger Commonwealth School of Medicine, Scranton, PA"
      ],
      "name": "C Neal"
    },
    {
      "affiliations": [
        "Neurosurgery, Geisinger Health, Danville, PA"
      ],
      "name": "O Goren"
    },
    {
      "affiliations": [
        "Geisinger, Wilkes-Barre, PA"
      ],
      "name": "C Schirmer"
    },
    {
      "affiliations": [
        "Geisinger, Danville, PA"
      ],
      "name": "P Hendrix"
    }
  ],
  "title": "E-144 Comparative performance of NIHSS-derived metrics for functional outcomes and mortality after acute large vessel occlusion stroke",
  "uid": "79f79b8a-9fc0-50d8-8038-e4771025b1b8"
}