Abstract
BACKGROUND: The use of anticoagulants and antiplatelet therapies is associated with a higher risk of bleeding in atrial fibrillation (AF) patients after percutaneous coronary intervention, especially after stent implantation. However, no accurate bleeding risk prediction tool has been developed for these patients. The aim of this study was thus to establish a bleeding risk prediction model (predictive nomogram) for patients with AF after stent implantation. METHODS: Construction of the predictive nomogram was based on a retrospective study, which enrolled 943 AF patients who underwent drug-eluting stent implantation between May 2012 and September 2016. A range of factors, including demographics, comorbidities, medication strategies, arterial access, and laboratory tests, were collected as baseline data. The least absolute shrinkage and selection operator (LASSO) and multivariate logistic regression analysis were used to identify the key clinical features for construction of the predictive nomogram. The concordance index (C-index) and internal validation were used to evaluate the efficacy of the nomogram. RESULTS: Of the 943 AF patients that underwent stent implantation, the occurrence of bleeding events was 8.2% (77 out of 943). Key predictors included the number of antiplatelet drugs, peptic ulcer, cerebral infarction, type 2 diabetes, thrombocytopenia, anemia, prior myocardial infarction, sex (male), use of anticoagulant drugs, liver dysfunction, hypertension, and acute myocardial infarction. These predictors were used to construct the nomogram. The C-index for the prediction of bleeding risk by the nomogram was 0.841 (95% CI: 0.79-0.89), which indicated good discrimination and calibration. The C-index of internal validation was 0.795, which demonstrated good efficacy of the model. CONCLUSIONS: This study suggests that our novel nomogram can accurately predict bleeding risk in AF patients after stent implantation during hospitalization, thereby helping to avoid complications. The nomogram may also be helpful for the creation of individualized post-discharge medication strategies.