Risk scores for predicting dysphagia in critically ill patients after cardiac surgery.

ABSTRACT

BACKGROUND: This study aimed at developing and validating a scoring model to stratify critically ill patients after cardiac surgery based on risk for dysphagia, a common but often neglected complication. METHODS: Data were prospectively collected and analyzed from January 2016 to June 2017 from 395 consecutive post cardiac surgery patients at the cardiac care unit (CCU) at a single center; 103 (26.1%) developed dysphagia. Univariate and multivariate logistic analyses were used to identify independent predictors for dysphagia. The survival nomogram was developed on the basis of a multivariable Cox model, which allowed us to obtain survival probability estimations. The predictive performance of the nomogram was verified for discrimination and calibration. Areas under receiver operating characteristic curve analysis were used to illustrate and evaluate the diagnostic performance of the novel model. RESULTS: The final novel scoring model, named SSG-OD, consists of three independent factors gastric intubation (OR = 1.024, 95% CI 1.015-1.033), sedative drug use duration (OR = 1.031, 95% CI 1.001-1.063) and stroke or not (OR = 6.182, 95% CI 3.028-12.617). SSG-OD identified patients at risk for dysphagia with sensitivity of 68.5% and specificity of 89.0% (OR = 0.833, 95% CI 0.782-0.884). The positive and negative likelihood ratios were 6.22 and 0.35. CONCLUSIONS: The novel SSG-OD scoring system to risk stratify CCU patients for dysphagia is an easy-to-use bedside prognostication aid with good predictive performance and the potential to reduce aspiration incidence and accelerate recovery.