The predictive dysphagia score (PreDyScore) in the short- and medium-term post-stroke: a putative tool in PEG indication

Purpose: We performed an evaluation of dysphagia in an unselected series of strokes to identify factors causing persisting dysphagia at 1 month after onset and to formulate a predictive score. Methods: We evaluated the association between dysphagia and clinical aspects (univariate analysis) at the 7th and 30th days after admission. We performed a multivariate logistic regression at the 30th day on the factors that were significant. We computed a simple score for predicting persistent dysphagia. Results: We recruited 249 patients. At the 7th day, 94 patients were dysphagic (37.75%). Factors associated with dysphagia included TACI (OR 3.85), mRS ≥ 3 (OR 4.45), malnutrition (OR 2.69), and BMI ≥ 20 (OR 0.52). At the 30th day, 217 patients remained in the study, and dysphagia persisted in 75 (36.76%). The factors that were associated with dysphagia were age > 74 years (OR 1.99), TACI (OR 5.82), mRS score ≥ 3 (OR 4.31), malnutrition (OR 3.27), and BMI ≥ 20 (OR 0.45). The multivariate analysis indicated that mRS ≥ 3 (OR 1.80) and BMI ≥ 20 (OR 0.45) remained significantly associated with dysphagia. The best correlation with dysphagia was the sum of mRS and the reciprocal of the BMI multiplied by 100 ((mRS + 1 ∕ BMI) × 100). We named this score PreDyScore that ranged between 3.7 and 10.47. Using < 6 and > 8 as cutoffs, the sensitivity was 67.03%, and the specificity 95.65%. Conclusion: BMI < 20 and mRS ≥ 3 are easily measurable bedside predictive factors of persistent dysphagia. PreDyScore showed good sensitivity and very good specificity and enables the prediction of persistent dysphagia with great accuracy in any clinical setting.