Incidence and predictors of bleeding complications after percutaneous coronary intervention.

BACKGROUND

Bleeding complications remain one of the most important challenges in percutaneous coronary intervention (PCI), particularly in Asians who are known to be vulnerable to the use of antiplatelets or anticoagulants. However, the incidence and predictors of bleeding complications after PCI have not been thoroughly investigated in Japan.

METHODS

We studied 13,075 consecutive patients in a Japanese multicenter PCI registry (Japan Cardiovascular Database - Keio interhospital Cardiovascular Study: JCD-KiCS) from September 2008 to March 2014. Multivariate logistic regression analysis was performed to investigate independent predictors of bleeding complications, and to create three risk prediction models for bleeding events. Model 1 included patients' characteristics alone. In model 2, we added patients' clinical presentation. Model 3 included covariates in model 2 along with angiographic and technical factors. Model discrimination was assessed using the area under the receiver operating curve (AUC).

RESULTS

Overall, bleeding complications, according to the pre-specified US National Cardiovascular Data Registry criteria, were observed in 402 patients (3.1%). Independent predictors of bleeding complications included age, female gender, previous PCI, previous heart failure, hemodialysis (variables included in model 1), ST-elevation and non-ST-elevation myocardial infarction, cardiogenic shock (added in model 2), transradial intervention, use of intra-aortic balloon pumping or a rotablator, and PCI for chronic total occlusion (added in model 3). Above all, previous PCI and transradial intervention were inverse predictors of bleeding. The predictability of the risk models improved as the number of variables increased, with AUC of 0.667, 0.747, and 0.791 for models 1, 2, and 3, respectively.

CONCLUSIONS

The incidence of bleeding complications among Japanese PCI patients was approximately 3% in standard nomenclature, which is equivalent to that of other international registries. Patients' characteristics, clinical presentation, and angiographic and technical factors all independently contributed to its prediction.