Association of Aortic Valve Disease Associated With Bleeding Arteriovenous Malformations of the Gastrointestinal Tract: Data From the Nationwide Inpatient Sample

Document Type


Publication Date



Introduction: Since the first descriptions of a correlation between Aortic Stenosis (AS) and Arteriovenous malformations (AVM) by Heyde in 1958, controversy has persisted about whether these diseases were truly associated or coincidental findings of older age. More recently, identification of decreases in large von Willebrand Factor (VWF) multimers which inversely correlated with aortic valve gradient and the observation that AVM bleeds decreased after valve repair led researchers to hypothesize that shear stress may be a potential mechanism. If this theory is correct, we postulated that a large dataset would show a significantly higher prevalence of bleeding AVMs in patients with high-gradient aortic valve disease as compared to low-gradient mitral valve disease.

Methods: We used the 2011 Nationwide Inpatient Sample (NIS) database to identify hospitalizations in patients with bleeding intestinal arteriovenous malformations (AVM) (International Classification of Diseases, 9th revision, Clinical Modification (ICD-9-CM) codes 537.83 and 569.85). ICD-9-CM codes were used to identify patients with aortic valve disease, mitral valve disease and known risk factors for AVM (including diagnosed von Willebrand disease, left ventricular assist device, chronic kidney disease). Univariate and multivariate logistic regression were used to determine the odds of association between the valvular diseases and AVM. Data analysis was done using STATA version 13.0 (College Station, TX).

Results: A total of 32,079 AVM related hospitalizations were identified of which 7.02% (n=2,253) cases had co-existent aortic valve disease. The unadjusted odds of aortic valve disease in association with bleeding AVMs versus those without bleeding AVMs was 4.95 (95% CI 4.43-5.54, p < 0.001). The association of AVM with mitral valve disease was not significant (OR: 1.56 (95% CI 0.59-4.14, p=0.38) (Table 1). When adjusted for age and known risk factors, the odds of aortic valve disease in bleeding AVMs was still significant (odds ratio (OR): 2.37, 95% CI 2.10-2.66, p < 0.001) (Table 2).

Publication Title

American College of Gastroenterology Annual Meeting

Open Access