Predicting Outcomes After Acute Reperfusion Therapy for Basilar Artery Occlusion.

2020 
BACKGROUND Basilar artery occlusion (BAO) leads to high rates of morbidity and mortality, despite successful recanalization. The discordance between flow restoration and long-term functional status clouds clinical decision making regarding further aggressive care. We sought to develop and validate a practical, prognostic tool for the prediction of 3-month favorable outcome after acute reperfusion therapy for BAO. METHODS This retrospective, multicenter, observational study was conducted at four high-volume stroke centers in the U.S. and Europe. Multivariate regression analysis was performed to identify predictors of favorable outcome (90-day mRS scores 0-2) and derive a clinically applicable prognostic model (the Pittsburgh Outcomes after Stroke Thrombectomy-VertebroBasilar (POST-VB) score). The POST-VB score was evaluated and internally validated apropos of calibration and discriminatory ability. External validity was assessed in patient cohorts at 3 separate centers. RESULTS In the derivation cohort of 59 patients, independent predictors of favorable outcome included smaller brainstem infarct volume on post-procedure MRI (p<0.01) and younger age (p=0.01). The POST-VB score was calculated as: age + (10 × brainstem infarct volume). The POST-VB score demonstrated excellent discriminatory ability (AUC=0.91) and adequate calibration (p=0.88) in the derivation cohort (center A). POST-VB performed equally well across the 3 external validation cohorts (center B, AUC=0.89; center C, AUC=0.78; center D, AUC=0.80). Overall, a POST-VB score <49 was associated with an 88% likelihood of favorable outcome, as compared to 4% with a score ≥125. CONCLUSIONS The POST-VB score effectively predicts 3-month functional outcome following acute reperfusion therapy for BAO and may aid in guiding post-procedural care.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    23
    References
    6
    Citations
    NaN
    KQI
    []