Validation of Prognostic Scales for Functional Outcome in Ischemic Stroke Patients Treated with Intravenous Thrombolysis in a Rural Setting

• Published in: Cerebrovascular diseases extra 2024-10, p.1
• Main Authors: Loggini, Andrea, Saleh Velez, Faddi G, Hornik, Jonatan, Henson, Jessie, Wesler, Julie, Hornik, Alejandro
• Format: Article
• Language: English
• ISSN: 1664-5456

Background Early prediction of functional outcome after rtPA helps clinicians in prognostic conversations with stroke patients and their families. Three prognostic tools have been developed in this regard: DRAGON, MRI-DRAGON, and S-TPI scales. These tools, all performing with a comparable accuracy, have been internally and externally validated in tertiary care centers. However, their performance in rural areas remains uncertain. This study addresses this gap in the literature by evaluating the effectiveness of those prognostic tools in stroke patients treated in a rural area of the Midwest. Methods We conducted a retrospective study of stroke patients treated with thrombolytics at Southern Illinois Healthcare Stroke Network from July 2017 to June 2024. Data on demographics, clinical presentations, laboratory values, neuroimaging, and stroke metrics were collected. modified Rankin Scale (mRS) at one month, classified into good (mRS ≤2) and poor (mRS ≥5) outcomes was noted. DRAGON and MRI-DRAGON scores were calculated. S-TPI model was built. Area under the receiver operating characteristic curve (AUC) with its 95% confidence interval was calculated for each prognostic model. Results A total of 279 patients were included in this study. Of those, 43% (n=119) were male. Median age (IQR) was 69 (57-80) years. NIHSS at presentation (IQR) was 7 (4-13). 12% of the cohort (n=34) had posterior circulation stroke. At one month, 66% of patients (n=185) had mRS≤2, whereas 14% of patients (n=39) mRS≥5. MRI-DRAGON showed the highest accuracy in predicting both good (AUC=0.86, 95%CI:0.81-0.90) and poor outcomes (AUC=0.84, 95%CI:0.76-0.91). DRAGON also demonstrated high accuracy for good (AUC=0.85, 95%CI:0.80 -0.89) and poor (AUC=0.82, 95%CI:0.75-0.90) outcomes. Conversely, in our population, the S-TPI model had the lowest accuracy for good (AUC=0.56, 95%CI:0.49-0.63) and poor (AUC=0.68, 95%CI:0.61-0.76) outcomes. Conclusions Among the available grading scores, MRI-DRAGON score can be considered the more accurate short-term prognostic tool for stroke patients treated with rtPA in the rural setting.