Abstract 19592: Predictors of Bleeding Complications in Patients Undergoing Percutaneous Coronary Intervention in an Australian Cohort
Background: Extensive studies have been undertaken to define clinical predictors of bleeding events in patients undergoing percutaneous coronary intervention (PCI). Few studies have included a large cohort of ‘real world’ patients undergoing PCI via a radial approach.
Methods: The Coronary Angiogram Database of South Australia (CADOSA) captures all PCI procedures performed in each public hospital in the state of South Australia (population 1.675 million). CADOSA utilizes identical data specifications to CathPCI® Registry. Thus a significant PCI-related bleeding event was defined as blood loss from any location (percutaneous entry site, retroperitoneal, gastrointestinal, genitourinary) within 72 hours of procedure, resulting in a fall in haemoglobin > 3.0g/dL, transfusion, or prolonged hospital admission. Clinical determinants of PCI-related bleeding events were identified in a multivariable logistic regression model.
Results: In 2012, there were 1,409 PCI procedures performed, with 80% undertaken for acute coronary syndrome, 56% via radial access, 14% using bivalirudin/enoxaparin and 5% utilizing novel anti-platelet agents (prasugrel 5%, ticagrelor 0.1%). Of these procedures, 1.7% had significant bleeding events. Univariate determinants of bleeding included increasing age, female, current smoker, prior PCI, ST elevation MI presentation, femoral access, cardiogenic shock, intra-aortic balloon pump insertion and administration of clopidogrel, glycoprotein IIb/IIIa inhibitors, verapamil or metaraminol prior/during procedure. The final model identified the following clinical determinants as significant independent predictors of bleeding with a C-index of 0.87: age over 80 years (OR=10.7; 3.4-34.0, p<0.01), current smoker (OR=5.7; 2.0-16.5, p<0.01), the use of glycoprotein IIb/IIIa inhibitor (OR=3.9; 1.4-10.6, p<0.01), use of clopidogrel (2.7; 1.0-7.1, p<0.05) and femoral access site (OR=10.6; 3.1-37.1, p<0.01).
Conclusion: In a ‘real world’, representative PCI cohort where radial access approach is frequently used, there was a low prevalence of bleeding events. In addition to access site, several distinctive baseline clinical features along with administration of blood thinning agents predicted bleeding.
- Percutaneous coronary intervention (PCI)
- Glycoprotein iib/iiia platelet inhibitors
- Risk factors
Author Disclosures: R. Tavella: None. M. Worthley: None. M. Arstall: None. C. Zeitz: None. J. Beltrame: Speakers Bureau; Modest; Servier Laboratories.
- © 2014 by American Heart Association, Inc.