Incidence, predictors, and prognostic implications of bleeding and blood transfusion following percutaneous coronary interventions

https://doi.org/10.1016/S0002-9149(03)00972-XGet rights and content

Abstract

Bleeding related to percutaneous coronary intervention (PCI) occurs relatively frequently. We retrospectively investigated the incidence, predictors, and prognostic impact of periprocedural bleeding and transfusion in 10,974 patients who underwent PCI. Bleeding definitions were based on Thrombolysis In Myocardial Infarction (TIMI) criteria: (1) major bleeding (n = 588; 5.4%): if patients had a hemorrhagic stroke or if hematocrit decreased >15 points or by 10 to 15 points with clinical bleeding; (2) minor bleeding (n = 1,394; 12.7%): if hematocrit decreased <10 points with clinical bleeding or by 10 to 15 points without clinical bleeding; and (3) no bleeding (n = 8,992; 81.9%): if hematocrit decreased <10 points without clinical bleeding. Patients with major bleeding were older than patients with minor or no bleeding (67.8 ± 11 vs 65.9 ± 11 vs 63.6 ± 11 years, respectively; p <0.001) and more often experienced intraprocedural complications, such as emergency use of an intra-aortic balloon pump (13.6% vs 6.5% vs 2.3%, respectively; p <0.001). Multivariate logistic regression analysis identified the use of an intra-aortic balloon pump (odds ratio [OR] 3.0, p <0.0001), procedural hypotension (OR 2.9, p <0.001), and age >80 years (OR 1.9 compared with age <50 years, p = 0.001) as the strongest predictors for major bleeding. Patients who had major bleeding had higher in-hospital and 1-year mortality compared with patients with minor or no bleeding. Bleeding was an independent predictor of in-hospital death. Thus, periprocedural major bleeding occurs relatively frequently and is associated with adverse outcomes. Patients >80 years of age who experience intraprocedural complications are at particularly high risk.

Section snippets

Study population

A retrospective analysis of our cardiac catheterization database identified 11,405 patients who underwent coronary intervention between 1991 and 2000 (431 patients who underwent coronary artery bypass grafting and PCI during the same admission were excluded). The final study population consisted of 10,974 patients who underwent 12,029 PCIs.

Study definitions and follow-up data

All data were confirmed by independent hospital chart review. Bleeding definitions were based on Thrombolysis In Myocardial Infarction (TIMI) criteria13 and

Baseline demographics

Major bleeding was observed in 588 patients (5.4%), minor bleeding in 1,394 patients (12.7%), and no bleeding in 8,992 patients (81.9%). Patient demographics are listed in Table 1. In those patients with acute myocardial infarction, thrombolytics were not significantly associated with bleeding. The proportion of women experiencing major bleeding (228 of 3,307; 6.9%) was 1.5 times higher than the proportion of men who developed bleeding (360 of 7,635; 4.7%). There was a downward trend in the

Discussion

The main findings of the present study of the 10,974 unselected consecutive patients who underwent PCI are: (1) periprocedural bleeding is associated with adverse in-hospital and 1-year outcomes; (2) IABP use, procedural hypotension, age >80 years and chronic renal insufficiency are strongly associated with bleeding; and (3) the need for a blood transfusion, independent of the bleeding episode, is associated with increased in-hospital and 1-year mortality. Previous reports of bleeding during

References (25)

Cited by (0)

View full text