Abstract: Objective To investigate the relationship between red blood cell distribution width(RDW), platelet count(PLT), hematocrit(HCT) and the risk degree of acute nonvariceal upper gastrointestinal bleeding(ANVUGIB) and the effectiveness of their combination in the prediction of prognosis. 
Methods A total of 300 ANVUGIB patients admitted to our hospital were selected and divided into death group(n=26) and survival group(n=274) according to the prognosis during hospitalization. The general data, RDW, PLT, and HCT levels in the two groups were compared. Pearson analysis was used to analyze the correlation between RDW, PLT, HCT and Glasgow-Blatchford scoring system(GBS), and partial correlation was used to further analyze the relationship between each index and GBS. Receiver operating characteristic curve(ROC) and the area under ROC curve(AUC) were used to analyze the effectiveness of RDW, PLT, HCT and their combination in predicting prognosis. 
Results The GBS score of the death group was higher than that of the survival group(P＜0.05), and the difference in the distribution of blood loss between the two groups was statistically significant(P＜0.05). The RDW of the death group was higher than that of the survival group, while the PLT and HCT were lower than those of the survival group(P＜0.05). RDW was positively correlated with GBS score(r=0.601, P＜0.001), PLT(r=-0.592, P＜0.001), and HCT(r=-0.697, P＜0.001), and were negatively correlated with GBS score. After controlling the bleeding volume, RDW, PLT and HCT were still correlated with GBS score(P＜0.05). The AUC of RDW+PLT+HCT in predicting prognosis was 0.900, which was higher than that predicted by each index alone. 
Conclusion RDW, PLT, and HCT are all related to the risk classification and prognosis of patients with ANVUGIB. Combined detection can improve the prognostic value and provide a basis for timely, effective and individualized treatment. 

Key words: gastrointestinal hemorrhage, red blood cell distribution width, platelet count, hematocrit