By Michael Crawford, MD, Editor

SYNOPSIS: In a recent analysis, abnormalities of von Willebrand factor were much more common in dysfunctional prostheses and were associated with gastrointestinal bleeding and angiodysplasia.

SOURCES: Blackshear JL, McRee CW, Safford RE, et al. von Willebrand factor abnormalities and Heyde’s syndrome in dysfunctional heart valve prostheses. JAMA Cardiol 2016;1:198-204.

Hillegass WB, Limdi NA. Valvular heart disease and acquired type 2A von Willebrand syndrome — The “hemostatic” Waring Blender syndrome. JAMA Cardiol 2016;1:205-206.

The association of severe aortic stenosis and gastrointestinal (GI) bleeding due to angiodysplasia (Heyde’s syndrome), is thought to be due to intravascular shear-induced changes in von Willebrand factor (VWF). Sporadic reports of similar VWF abnormalities and GI bleeding in other dysfunctional valves led researchers to prospectively study 136 patients: 26 presenting with normally functioning surgical or transcatheter aortic prosthetic valves, 24 containing dysfunctional aortic prostheses, 36 enjoying normally functioning mitral valve prostheses or repairs; 19 suffering from dysfunctional mitral prostheses or repairs; and 31 patients featuring native isolated aortic regurgitation. A dysfunctional prosthesis was one with at-least-moderate stenosis or regurgitation. Several tests of VWF and platelet function were measured. In patients with native isolated aortic regurgitation (AR), abnormal VWF multimers were present in 77% of those with moderate-to-severe AR and regurgitant volume correlated with platelet dysfunction (r = 0.74; P < 0.001). Abnormal VWF multimers were more common in patients with dysfunctional aortic prostheses compared to those with normal function (83% vs. 4%; P < 0.001). Researchers observed similar findings with dysfunctional vs. normally functioning mitral valves (74% vs. 6%;
P < 0.001). GI bleeding occurred in two of 21 patients with at-least-moderate native AR, but in six of 24 patients with dysfunctional aortic prosthesis and five of 19 with mitral prosthesis dysfunction. Among those with aortic prostheses dysfunction and GI bleeding, five of six demonstrated angiodysplasia. In those with mitral prosthesis dysfunction with GI bleeding, three of five experienced angiodysplasia and the other two did not undergo endoscopy. The authors concluded that in patients with dysfunctional aortic and mitral valve prostheses, acquired abnormalities in VWF are associated with GI bleeding and angiodysplasia. They suggested that measuring VWF multimers may provide additional diagnostic and prognostic information in patients with dysfunctional valve prostheses.


Frank hemolysis with schistocytes on the blood smear due to prosthetic valve dysfunction is rare, but VWF syndrome with angiodysplasia and GI bleeding is common. This syndrome has been described in moderate-to-severe aortic stenosis, severe AR, severe mitral regurgitation, hypertrophic obstructive cardiomyopathy, various congenital heart defects, and left ventricular assist devices. VWF is a large, multimerized protein that contains binding sites for collagen, platelet glycoprotein, and factor VIII. When the arterial wall is breached, VWF adheres to the site of injury, recruits platelets, and binds factor VIII, leading to the blood clotting cascade. Increased blood shear forces as seen in the disorders listed above results in cleavage of the multimers and a reduction in the number of multimers. Too few multimers leads to bleeding, especially at high shear points such as angiodysplasia in the gut. Too much VWF leads to abnormal clotting.

This paper analyzed VWF function parameters and clinical events in a series of patients with native and prosthetic valve disease and showed that abnormalities of VWF are found in about 80% of dysfunctional aortic valve prostheses, whether placed surgically or by transcatheter; dysfunctional mitral valve repairs and prostheses; and moderate-to-severe native valve AR. More importantly, about one-fourth of prosthetic valve dysfunction patients experienced GI bleeding that was almost always associated with angiodysplasia. GI bleeding occurred in only 10% of the moderate-to-severe native AR patients, probably because shear forces are lower in AR due to lower pressure gradients in AR as compared to aortic stenosis or mitral regurgitation.

The authors suggested two potential clinical uses of this information. First, VWF measures could be used as a biomarker for the severity of native valve AR and prosthetic valve dysfunction. Interestingly, they also measured brain natriuretic peptide levels in this study and they were not discriminatory. Second, reduced VMF multimers may identify patients at risk for GI bleeding. Unfortunately, there is only one way to fix increased shear stress: fix or replace the valve. This is not always a straightforward decision for someone who just underwent valve surgery. Also, many valve disease patients are on warfarin, especially if they also suffer from atrial fibrillation. Bleeding often is mistakenly attributed to anticoagulation. Measure VWF in any of these patients with bleeding.

There are several limitations to this study. First, the incidence of VWF abnormalities and bleeding may be exaggerated due to referral bias. Also, there were only four patients in this series with severe native AR. In addition, due to the relatively low number of patients overall, information about specific prosthetic valve malfunctions and their outcomes, especially after corrective procedures, is too small to analyze meaningfully. Patients in the series featured patient-prosthesis mismatches, prosthetic stenosis, and perivalvular leaks. Despite these limitations, this study clearly identified the potential importance of measuring VWF function in patients with suspected left-sided valve prosthesis dysfunction and those with abnormal bleeding. At this point, there are no agreed-upon cutpoints for decision-making, but the information may push the need for a corrective procedure one way or another.