부산대학교 도서관

Background and Purpose:Post stroke dysphagia (PSD) is a common complication after stroke. The earlier prediction of PSD is essential for patient stratification for intensive rehabilitation, but there was no scoring system for predicting PSD. We aimed to develop the PSD prediction score.Methods:We examined consecutive patients with acute ischemic stroke between 2018-2020. Patients with in-hospital-death were excluded. Dysphagia state was assessed using food oral intake scale (FOIS) score at hospital discharge. PSD was defined as FOIS 1-3, which represent tube-dependent nutrition. For the assessment of reproducibility of PSD score, patients were divided into derivation and validation cohort. Using the derivation cohort, PSD score was developed from associated factors to discriminate PSD by logistic regression analysis. Discriminative performance was analyzed by receiver operating curve (ROC) analysis. Reproducibility of PSD score was validated using the validation cohort.Results:Among 795 patients (median 77 years; male, 57%), 556 (70%) patients were assigned as derivation cohort and 239 (30%) patients were assigned as validation cohort. In the derivation cohort, older age (≥85 years), low BMI (≦18.5 kg/m2for aged <70 years and ≦ 20 kg/m2for aged ≥70 years), high NIHSS score (≥16), and low serum albumin (≦3.0 mg/dl) were associated with PSD (Table). PSD score were ranged from 0 to 6 from these factors (Table). Area under the curve (AUC) of PSD score for PSD was 0.83 (95% confidence interval [CI] 0.78-0.88) with an appropriate cut-off value of 2 (sensitivity, 69%; specificity, 88%). In the validation cohort, the AUC of PSD score for PSD was 0.78 (95% CI 0.70-0.86), which was not significantly different in the AUC of the derivation cohort (AUC; 0.83 vs 0.78, p=0.30).Conclusions:In acute ischemic stroke population, the discriminative performance of PSD score had acceptable for diagnosing PSD.