RESUMO
OBJECTIVES: Despite being the current standard of care, outcomes after endovascular thrombectomy (EVT) for acute ischemic stroke (AIS) remain highly variable. Though several scoring systems exist to predict outcomes in AIS, they were mainly developed to direct patient selection for treatment. Recognizing the integral role peri-procedural metrics play on outcome, our study aimed to develop a post-EVT prognostic score to predict 90-day functional dependency and death. MATERIALS AND METHODS: We included all eligible adult AIS patients treated with EVT at our institution from June 2016 to January 2020. Data was systematically collected via chart review including pre-, intra- and post-procedural variables. The outcome was modified Rankin score (mRS) at 90 days post-EVT where a poor outcome was defined as mRS 3-6: 3-5 for functional dependency and 6 for death. Model selection methods including stepwise and Lasso were evaluated via cross-validation where the final multivariable logistic regression model was chosen by optimizing the Area Under the Receiver Operating Characteristic Curve (ROC AUC). RESULTS: We included 224 patients (mean age: 65 years old, male: 55%, 90-day poor outcome: 60%). The final model achieved a median AUC of 0.84, IQR: (0.80, 0.87). A 7-point score, called Bronx Endovascular Thrombectomy (BET) score, was developed with more points indicating higher likelihood of 90-day poor outcome (0 point: ≤21% risk; 1-2: 24%; 3: 61%; 4: 86%; 5: 96%; 6-7: ≥99%). One point was awarded for the following variables: current smoker, diabetic, general anesthesia received, puncture to perfusion time ≥45 minutes, and Thrombolysis in Cerebral Infarction (TICI) score <3. Two points were awarded for a post-EVT National Institute of Health Stroke scale (NIHSS) of ≥10. CONCLUSION: Incorporating peri-procedural data we developed the competitive BET score predicting 90-day functional dependency and death, which may help providers, patients and caregivers manage expectations and organize early rehabilitative services.
