The risk of stroke and death in patients with aortic and mitral valve endocarditis.

Journal Article (Journal Article)

BACKGROUND: Previous studies have generated inconsistent results when attempting to define predictors of stroke and death in patients with endocarditis. We sought to examine the relationship between vegetation 2-dimensional size and stroke in those with infective endocarditis (IE) and to identify differences between aortic valve (AV) and mitral valve (MV) IE with regard to clinical characteristics, echocardiographic findings, stroke, and death. METHODS: We used the Duke Endocarditis Database to examine 145 episodes of definite IE involving the AV, n = 62, or MV, n = 83. A logistic regression model was developed to analyze important variables in predicting stroke, and a Cox proportional hazards model was used in predicting mortality. RESULTS: The mitral valve was infected in 57% of the cases. Vegetations were more commonly detected in patients with MV IE (92.8% vs 66.1%, P =.001) and these MV vegetations were significantly larger (P <.05). Thirty-four of 145 episodes (23.4%) were complicated by stroke. MV IE was associated with a greater stroke rate, 32.5% versus 11.3% (P =.003). Strokes tended to occur early in the course of illness, particularly in MV IE. In the multivariable model, the independent predictors of stroke were MV IE (P =.04) and vegetation length (P =.03). Independent predictors of 1-year mortality were age (P =.02) and vegetation area (P =.048). CONCLUSION: Stroke is more common in patients with MV IE. Vegetation 2-dimensional size and characteristics are important predictors of stroke and mortality. These findings may lead to predictive models that allow physicians to identify high-risk patients who need aggressive treatment strategies to prevent long-term morbidity and mortality.

Full Text

Duke Authors

Cited Authors

  • Cabell, CH; Pond, KK; Peterson, GE; Durack, DT; Corey, GR; Anderson, DJ; Ryan, T; Lukes, AS; Sexton, DJ

Published Date

  • July 2001

Published In

Volume / Issue

  • 142 / 1

Start / End Page

  • 75 - 80

PubMed ID

  • 11431660

International Standard Serial Number (ISSN)

  • 0002-8703

Digital Object Identifier (DOI)

  • 10.1067/mhj.2001.115790


  • eng

Conference Location

  • United States