• Shenyan Gu, Yuqin Wang, Kaifu Ke, Xin Tong, Jiahui Gu, and Yuanyuan Zhang.
• Department of Neurology, Affiliated Hospital of Nantong University, Medical School of Nantong University, Nantong, China.
• Curr Med Res Opin. 2022 Nov 1; 38 (11): 192319331923-1933.

BackgroundRichmond agitation-sedation scale (RASS) is a simple and widely used tool for evaluating sedation and agitation in adult ICU patients. Early deep sedation has been shown to be an important independent predictor of death, however, studies on the role of RASS in the prognostic assessment of neurocritical patients are lacking. The purpose of this study was to investigate the relationship between RASS and in-hospital mortality in neurocritical patients, and to develop and validate an effective predictive model based on this.MethodsThis was a retrospective study of neurocritical patients from a large clinical database. A total of 2651 patients were collected, including general demographic characteristics, past medical history, biochemical test data and physical examination within 24 h of admission, and related medical records. Univariate and multivariate logistic regression analyses were used to screen out significant variables. Finally, 11 significant predictors were included into the logistic regression to establish the nomogram.ResultsThe area under the curve (AUC) of the nomogram was 0.9087(0.8950-0.9224) and the corrected c index was 0.9043, which gave the model better discriminatory ability compared with critical care related scales, such as SOFA and SAPSII scores. Besides, tools including calibration curve, decision curve analysis (DCA), and clinical impact curve (CIC) were used to verify that the model had good discrimination, calibration, and clinical applicability.ConclusionsRASS score was an independent prognostic predictor of in-hospital death in neurocritical patients, and patients who are deeply sedated have a worse prognosis. RASS-related nomogram could be applied to predict the prognosis of neurocritical patients and to take effective intervention measures in early stage.

10 keywords...

hide…