Comparison of predictive value of risk scores for gastrointestinal bleeding in antiplatelet therapy

Gastrointestinal bleeding (GIB) is the most common serious bleeding complication of antiplatelet therapy. The bleeding risk score (BRS) of GIB may help to determine the risk of bleeding, and provides a reference for the formulation of antiplatelet therapy regimen in clinical practice, but we found that no specific risk scores are available in East Asian patients. This study analyzed patients who were administered antiplatelet therapy from May 2015 to December 2018 in two medical centers. Patient's baseline data were obtained. We assessed four BRSs (New Score, RIETE Score, Cuschieri Score, de Groot Score) and compared them using the area under the receiver operating characteristic curve (AUC). The 4,052 patients enrolled in this study had an average age of 69.6 ± 10.8 years, and 65.9% of them were male. Among the 4,052 patients included, 171 patients experienced GIB within 6 months of follow-up. In the study population, the AUCs for the New, RIETE, Cuschieri, and de Groot scores were 0.673 (95% confidence interval (CI) 0.616-0.729, P < .001), 0.742 (95% CI 0.690-0.794, P < .001), 0.598 (95% CI 0.537-0.659, P = .002), and 0.875 (95% CI 0.839-0.912, P < .001), respectively. After validation, the de Groot Score has better performance. Among the four scores, the de Groot Score might be more suitable for helping Chinese clinicians to predict the risk of GIB in patients taking antiplatelet drugs, and reduce GIB events.