Evaluating the Relationship Between Gastrointestinal Bleeding and Valvular Heart Disease: A Systematic Review of Clinical Studies

Abstract

Background: Gastrointestinal angiodysplasia is a significant vascular anomaly characterized by dilated, tortuous blood vessels in the gastrointestinal tract. The current literature extensively documents the association between angiodysplasia and aortic stenosis, known as Heyde syndrome, characterized by the triad of aortic stenosis, GIB, and acquired von Willebrand syndrome. However, other valvular diseases, including mitral and tricuspid regurgitation, have also been implicated. This comprehensive systematic review aims to investigate the spectrum of valvular abnormalities, exploring the intricate mechanisms by which they contribute to gastrointestinal bleeding. Furthermore, it will evaluate the available surgical and nonsurgical treatment modalities, assessing their efficacy in mitigating the incidence of such bleeding. Methods: A comprehensive search of the Pubmed/MEDLINE database was conducted to identify relevant studies to retrieve relevant articles from August 2014 to August 2024. A combination of Medical Subject Heading (MeSH) terms and text words related to cardiac valvular diseases and GIB were used. MeSH terms included “gastrointestinal bleeding”, “heart valve diseases”, “hematochezia”, “heart valve prosthesis”, “bioprosthesis”, “native valve diseases”, and “mechanical valve”. Results: Forty-five papers met the inclusion criteria. Twenty-seven studies covered GIB in aortic valve disease, ten on mitral valve disease, two on tricuspid valve disease, and six on multiple valves. Conclusions: This systematic review demonstrates the association between angiodysplasia and aortic stenosis and highlights mitral regurgitation and tricuspid regurgitation as potential etiologies. Definitive management with valvuloplasty or valve replacement is vital to preventing the onset or recurrence of GIB in patients with valvular disease.

1. Introduction

Gastrointestinal angiodysplasia is a significant vascular anomaly characterized by dilated, tortuous blood vessels in the gastrointestinal tract. This condition is most commonly found in the cecum and ascending colon but can occur throughout the GI tract. Angiodysplasia is a significant cause of both overt and occult gastrointestinal bleeding (GIB), particularly in older adults. The pathogenesis of angiodysplasia is not fully understood, but it is thought to involve chronic, intermittent obstruction of submucosal veins, leading to increased vascular endothelial growth factor (VEGF)-dependent proliferation and subsequent vessel dilation. Clinically, angiodysplasia can present with symptoms ranging from asymptomatic incidental findings to life-threatening GI bleeding. Diagnosis is often challenging due to the intermittent nature of bleeding and the inaccessibility of some lesions [1,2,3,4].

The current literature extensively documents the association between angiodysplasia and aortic stenosis, known as Heyde syndrome, characterized by the triad of aortic stenosis, GIB, and acquired von Willebrand syndrome. However, other valvular diseases have also been implicated, including mitral regurgitation and tricuspid regurgitation [5,6]. Bleeding in Heyde syndrome is thought to be multifactorial and leads to gastrointestinal angiodysplasia. Angiodysplastic vessels generate more shear stress than normal vessels, and the lack of functional von Willebrand factor (VWF) in aortic valve stenosis patients makes them prone to GIB [7].

The well-established association between valvular disease and an increased incidence of GIB underscores an essential area of current medical practice. This comprehensive study aims to investigate the spectrum of valvular abnormalities, exploring the intricate mechanisms by which they contribute to gastrointestinal bleeding. Furthermore, it will evaluate the available surgical and nonsurgical treatment modalities, assessing their efficacy in mitigating the incidence of such bleeding. By providing an in-depth analysis, this study seeks to enhance clinician understanding and inform better clinical practices to improve patient outcomes.

2. Methods

A comprehensive search of the Pubmed/MEDLINE database was conducted to identify relevant studies to retrieve relevant articles from August 2014 to August 2024. A combination of Medical Subject Heading (MeSH) terms and text words related to cardiac valvular diseases and GIB were used. MeSH terms included “gastrointestinal bleeding”, “heart valve diseases”, “hematochezia”, “heart valve prosthesis”, “bioprosthesis”, “native valve diseases”, and “mechanical valve”. Only English-language studies involving human subjects were included, and a range of study types, including prospective cohort studies, experimental studies, population studies, meta-analyses, retrospective cohort studies, clinical trials, and observational studies, were eligible for inclusion. Case reports and series were excluded unless at least five patients were included in the analysis. After screening and data extraction by two independent reviewers (Figure 1), the authors conducted a narrative synthesis of the studies according to the PRISMA guidelines, with the extracted data being summarized into tables for easy comparison and review.

3. Results

Two hundred eighty-six papers from PubMed/MEDLINE were reviewed. Forty-five met the inclusion criteria (Figure 1). Twenty-seven studies covered GIB in aortic valve disease, ten studies were on mitral valve disease, two on tricuspid valve disease, and six covered multiple valves. The studies are summarized in

Table 1. Summary of studies evaluating GIB in aortic valve disease, listed in order of appearance in text. Abbreviations: aOR, adjusted odds ratio; AS, aortic stenosis; AVR, aortic valve replacement; CABG, coronary artery bypass grafting; CI, confidence interval; GAVE, gastric antral vascular ectasia; LOS, length of stay; NNH, number needed to harm; RCT, randomized controlled trial; SAVR, surgical aortic valve replacement; TAVI, transcatheter aortic valve intervention; TEE, transesophageal.

Author, Year	Valvular Disorder	Study Type	Number of Patients	Outcomes Measured	Groups	Study Period	Notable Outcomes	Conclusions
Aldiabat et al., 2024 [8]	AS	Retrospective Observational	n = 85,090 patients with GAVE	Incidence of GIB due to GAVE	AS (n = 5315) vs. No AS (n = 79,775)	2016–2019	Patients with AS have a 2-fold increase in the risk of development of GAVE (aOR 2.08, 95% CI 1.94–2.22, p < 0.001). However, no difference in mortality was noted.	A strong association between AS and GAVE exists.
Hung et al., 2022 [9]	AS	Retrospective Observational	n = 78 uremic patients with persistent or recurrent bleeding of unknown origin after negative endoscopy undergoing push enteroscopy	Rebleeding rates and reasons	-	2010–2017	The rebleeding rate was 29.5%, with angiodysplasia as the most common etiology (74.6%) and the most common site being the jejunum (50.8%). Angiodysplasia and valvular heart disease were associated with a higher rebleeding rate (p < 0.05).	Jejunal angiodysplasia is a frequent etiology of unknown bleeding in uremic patients.
Waldschmidt et al., 2021 [10]	TAVI	Prospective Observational	n = 2548 TAVI patients with Heyde syndrome	Outcomes of pts with Heyde syndrome including bleeding complications and the recurrence of GIB	-	2008–2017	A history of GIB prior to TAVI was detected in 190 patients (7.5%). Among them, 47 patients were diagnosed with HS (1.8%). Heyde patients required blood transfusions more frequently compared to non-Heyde patients during index hospitalization (50.0% vs. 31.9%, p = 0.03). Recurrent GIB was detected in 39.8% of Heyde compared to 21.2% of non-Heyde patients one year after TAVI (p = 0.03). In patients diagnosed with HS and recurrent GIB after TAVI, the rate of residual ≥ mild paravalvular leakage was higher compared to those without recurrent bleeding (73.3% vs. 38.1%, p = 0.05).	Residual paravalvular leakage following TAVI may be associated with higher rates of recurrent GIB.
Jehangir et al., 2017 [11]	AS	Retrospective Observational	n = 32,079 patients with gastrointestinal angiodysplasia	Incidence of GIB and aortic valve disease	-	2011	A total of 7.02% of pts with angiodysplasia-related GIB had AS. The unadjusted odds of aortic valve disease associated with bleeding intestinal angioectasia versus those without bleeding angioectasia were 4.95 (95% CI: 4.43–5.54, p < 0.001) and 2.37 (95% CI 2.10–2.66, p < 0.001) after adjusting for age and known risk factors.	A strong association between aortic valve disease and bleeding intestinal angioectasia exists.
Tamura et al., 2015 [12]	AS	Prospective Observational	n = 31 severe AS patients	Analysis of the multimeric structure of VWF	-	2015	The loss of vWF multimer was detected in 67.7%. A total of 38.7% had Heyde syndrome. AVR was performed in 7 pts and resulted in a significant hemoglobin improvement (7.5 g/dL pre-op vs. 12.4 d/dL post-op, p < 0.0001).	The prevalence of acquired von Willebrand syndrome in patients with AS is high and should be considered in pts with AS who have anemia.
Selvam et al., 2020 [13]	AS	Prospective Cohort	n = 29 AVR patients	Plasma VWFAg, VWFAct, propeptide, collagen binding, multimers, factor VIII coagulant activity, ADAMTS13 activity before AVR, 3–5 days after, and 6 months after	n = 29 AVR patients vs. n = 10 controls	2020	AS patients exhibited quantitative and qualitative abnormalities of VWF including significantly increased VWF antigen, activity, and propeptide levels following surgery (p < 0.01). Increased high-molecular-weight VWF multimers were observed at all time points and, in particular, 3 to 5 days after surgery (mean 14% ± 6%) relative to before (mean 10% ± 4%).	Hemostatic changes in AS patients are present before valve replacement surgery and these persist long after surgery has occurred.
Jackson et al., 2014 [14]	AS	Systematic Review with Meta-analysis	n = 831 patients with gastrointestinal angiodysplasia from 22 studies	Cessation of bleeding	Endoscopic therapy (n = 623) vs. Hormonal therapy (n = 63) vs. Somatostatin analog (n = 72) vs. AVR (n = 73)	1970–2013	Hormonal therapy was not effective for bleeding cessation (OR 1.0, 95% CI 0.5–1.96). Endoscopic therapy was effective as initial therapy, but the pooled recurrence bleeding rate was 36% (95% CI 28–44%) over a mean (+/-s.d.) of 22+/-13 months. The event rate for rebleeding increased to 45% (95% CI 37–52%) when studies including only small-bowel GIADs were included (n = 341). The pooled OR of somatostatin analog therapy was 14.5 (95% CI: 5.9–36) for bleeding cessation. The event rate for rebleeding in those treated with AVR was 0.19 (95% CI: 0.11–0.30) over a mean follow-up period of 4 years postoperatively.	Endoscopic intervention may not offer enough therapeutic value as compared to AVR in pts with AS and GIB.
Caspar et al., 2015 [15]	AS	Prospective Observational	n = 49 patients with severe AS before and after TAVI	VWFAct, survival, bleeding complications	-	2012–2013	The mean aortic transvalvular gradient was negatively correlated with the levels of VWFAg (p < 0.05), VWF ristocetin cofactor activity (p = 0.006), and VWFCB (p = 0.005). One week after the procedure, a significant increase in VWFAg (3.32 vs. 2.29 IU/mL, p < 0.001), VWF ristocetin cofactor activity (2.98 vs. 1.86 IU/mL, p < 0.001), and VWFCB (3.16 vs. 2.16 IU/mL, p < 0.001) was observed.	Hemostasis parameters improve after TAVI, especially in those with pre-existing abnormalities.
Goltstein et al., 2022 [16]	AVR	Retrospective Observational	n = 70 TAVI patients with Heyde syndrome	Rates of GIB pre- and post-TAVI	-	2008–2020	The rate of GIB improved after TAVI after the first year (62% of patients, 95% CI, 50–74%). Bleeding episodes decreased from 3.2 (95% CI 2.5–4.2) to 1.6 (95% CI 1.2–2.2, p = 0.001). Hemoglobin levels increased from 10.3 (95% CI 10.0–10.8) to 11.3 (95% CI 10.8–11.6) g/dL (p = 0.007). Between 1 and 5 years after TAVI, 83% (95% CI 72–92%) no longer experienced gastrointestinal bleeding.	GIB rates improve following TAVI in patients with Heyde syndrome.
Tjahjadi et al., 2017 [17]	AS	Retrospective Cohort	n = 9562 severe AS patients	Prevalence of anemia	AVR (n = 1537) vs. CABG (n = 8025)	2002–2014	In the AVR group, 30.1% were anemic, compared (95% CI 27.9–32.5%) to 16.2% (95% CI 15.4–17.0%) in the CABG group. The adjusted marginal mean hemoglobin value was 135.6 g/L in AVR patients compared to 137.3 g/L in CABG patients.	The prevalence of anemia is higher in patients with AS who underwent AVR compared to those who undergo CABG.
Sugino et al., 2023 [18]	AS	Retrospective Observational	n = 325 patients who underwent TAVI with severe AS	Status of anemia and endoscopic features	-	2016–2019	The rates of moderate/severe anemia (hemoglobin < 11 g/dL) were 52%, and it was associated with significantly more angiodysplasia (38.3% vs. 7.7%, p < 0.0001) and active bleeding (23.4% vs. 0%; p < 0.01). Angiodysplasia was detected in 21 patients (stomach, n = 9; small intestine, n = 5, and colon, n = 10).	Angiodysplasia is associated with moderate/severe anemia in AS.
Stanger et al., 2017 [19]	TAVI	Retrospective Observational	n = 841 TAVI patients	Risk of severe upper GIB	-	2005–2014	The risk of upper GIB in TAVI patients was 2%, which increased 10-fold with triple antithrombotic therapy (11.8% vs. 1.0%) compared to those without any antithrombotic therapy. Endoscopy findings demonstrated five high-risk esophageal lesions including erosive esophageal ulcers, visible vessels at the GE junction, erosions at the distal esophagus, and an actively bleeding esophageal ring that had been intubated through by the TEE probe.	TAVI is associated with a moderate risk of severe upper GIB which increased significantly with triple antithrombotic therapy.
Manasrah et al., 2023 [20]	TAVI	Retrospective Observational	n = 320,353 TAVI patients	GIB and non-GIB cohorts in patients with TAVI during index hospitalization and at 30, 90, and 180 days after discharge	-	2016–2020	A total of 6193 patients were readmitted for GIB between 2016 and 2020, with higher readmission rates at 30, 90, and 180 days observed in those patients with GIB (aOR 6.35, 95% CI 5.37–7.52, p < 0.0001). GIB was also associated with higher in-hospital mortality (aOR 6.35, 95% CI 5.37–7.52, p < 0.0001).	Patients with TAVI who develop GIB have higher readmission and in-hospital mortality rates.
Zahid et al., 2022 [21]	TAVI	Retrospective Cohort	n = 216,023 TAVI patients	Incidence and predictors of GIB	GIB (n = 2188) vs. Propensity-matched controls (n = 2190)	2011–2018	Arteriovenous malformation was associated with the highest odds of having a gastrointestinal bleed (OR 24.8, 95% CI 17.13–35.92). Peptic ulcer disease was associated with an eightfold increase in the risk of bleeding (OR 8.74, 95% CI 6.69–11.43), followed closely by colorectal cancer (OR 7.89, 95% CI 5.33–11.70). Other comorbidities that were associated with higher propensity-matched rates of GIB were chronic kidney disease (OR 1.27, 95% CI 1.14–1.41), congestive heart failure (OR 1.18, 95% CI 1.06–1.32), liver disease (OR 1.83, 95% CI 1.53–2.19), end-stage renal disease (OR 2.08, 95% CI 1.75–2.47), atrial fibrillation (OR 1.63, 95% CI 1.49–1.78), and lung cancer (OR 2.80, 95% CI 1.77–4.41). Patients with GIB had higher propensity-matched rates of mortality than those without GIB (12.1% vs. 3.2%, p < 0.01). Patients with GIB had a higher median cost of stay (USD 68,779 vs. USD 46,995, p < 0.01) and a longer length of hospital stay (11 vs. 3 days, p < 0.01).	Health care use and mortality are higher in hospitalizations of TAVI with a GIB. Baseline comorbidities like peptic ulcer disease, chronic kidney disease, liver disease, atrial fibrillation, and colorectal cancer are significant predictors of this adverse event.
Laflamme et al., 2014 [22]	TAVI	Retrospective Observational	n = 18 severe AS patients who underwent TAVI	Clinical outcomes of TAVI with left axillary approach	-	2010–2012	The axillary approach has a higher procedural success rate and lower rates of complications, but in one case, upper GIB was seen.	The axillary approach may be an alternative approach for TAVI while noting a risk of UGIB.
Brown et al., 2022 [23]	AS	Retrospective Observational	n = 1192 patients with severe AS who underwent TAVI	Frequency of GIB	-	2011–2018	Of the 1192 patients who underwent TAVI, 164 developed severe GIB. A total of 130 (79.3%) had resolution of GIB after TAVI.	TAVI is associated with GIB improvements in patients with severe AS.
Yerasi et al., 2020 [24]	TAVI	Retrospective Observational	n = 49,742 TAVI patients	30-day readmission rates with next-day discharge	-	2012–2016	The rate of next-day discharges post-TAVI increased from 1.5% in 2012 to 12.2% in 2016, with similar readmission rates. In 2016, there was a significantly higher incidence of GIB.	The rates of next-day discharges and GIB in TAVI patients have increased from 2012 to 2016.
Iyengar et al., 2018 [25]	AS	Retrospective Observational	n = 353,370 patients with severe AS who underwent TAVI or SAVR	Incidence of late GIB	TAVI (n = 43,357) vs. SAVR (n = 310,013)	2011–2014	More pts with TAVI were re-hospitalized with GIB as compared to SAVR (aOR 1.54, 95% CI 1.38–1.71, p < 0.001).	TAVI is associated with more frequent readmissions for late GIB than SAVR.
Desai et al., 2019 [26]	AS	Retrospective Observational	n = 130,487 patients with AS hospitalized for GIB related to Heyde syndrome	Trends in all-cause in-hospital mortality	-	2007–2014	Hospitalizations (48 to 62 per 100,000) and all-cause inpatient mortality (3.7 to 4.54 per 100,000) increased compared to those with AS and without Heyde syndrome. All-cause mortality (6.9% vs. 4.1%), LOS (∼7.0 vs. 5.8 days), and hospitalization charges (USD 58,519.31 vs. USD 57,598.67) were higher in HS (p < 0.001). TAVI patients showed lower rates of stroke (1.7% vs.10.0%) and blood transfusion (1.7% vs. 11.7%), a shorter LOS (18.3 vs. 23.9 days; p < 0.001), and more routine discharges (21.7% vs. 14.8%, p = 0.01) compared to SAVR.	Higher rates of hospitalization and mortality were seen in patients with Heyde syndrome between 2007 and 2014. TAVI has improved stroke, transfusion rates, LOS, and routine discharges compared to SAVR.
Goltstein et al., 2023 [27]	AVR	Systematic Review with Meta-analysis	n = 1054 patients with AS who underwent TAVI or SAVR	Recovery of acquired von Willebrand syndrome within 24 h, 24 to 72 h, 3 to 21 days, and 4 weeks to 2 years after aortic valve AVR	-	Up to Oct 25th of 2022	AVR led to the recovery of vWF deficiency in 85.9% (95% CI, 79.1–90.7%; I2 = 26%) of patients within 24 h post-procedure, 89.5% (74.1–96.2%; I2 = 80%) between 24 and 72 h, 92.2% (84.0–96.3%; I2 = 58%) between 3 and 21 days, and 87.4% (67.2–95.9%; I2 = 84%) between 4 weeks and 2 years. AVR was also linked to a complete cessation of GIB in 72.8% (95% CI 62.2–81.3%; I2 = 59%) of patients, with follow-up periods ranging from a median of 12 to 108 months.	AVR is associated with rapid recovery of bleeding diathesis in Heyde syndrome.
Matsuura et al., 2023 [28]	AVR	Retrospective Observational	n = 1529 AVR patients	Primary: All-cause mortality Secondary: cerebral infarction, cerebral bleeding, GIB	Bioprosthetic valve (n = 837) vs. Mechanical valve (n = 692)	2010–2012	The Cox regression model showed no significant difference in long-term survival between the groups (mechanical valve: HR 0.895, 95% CI 0.719–1.113, p = 0.318). The incidences of cerebral infarction and prosthetic valve failure were similar between the groups. Cerebral bleeding (HR 2.143, 95% CI 1.125–4.080, p = 0.002) and GIB (HR 2.071, 95% CI 1.243–3.451, p = 0.0005) were more frequent in the mechanical valve group.	AVR with mechanical valves has a higher risk of GIB compared to bioprosthetic.
Kyto et al., 2019 [29]	AVR	Retrospective Observational	n = 4227 SAVR patients age > 70	10-year survival after operation	Mechanical valve (n = 296) vs. bioprosthetic valve (n = 3931)	2004–2014	The 10-year major bleeding rates were 37.0% with mechanical valves and 18.8% with biologic valves (HR, 1.77, 95% CI 1.25–2.49; p = 0.001; NNH = 7.4). GIB was significantly more frequent with mechanical valve prosthesis (26.5% vs. 8.9%; HR 2.63, 95% CI 1.63–4.23, p < 0.001).	Bioprostheses should be considered over mechanical prostheses in patients > 70 due to lower bleeding risks.
Rashvand et al., 2023 [30]	AVR	Retrospective Observational	n = 1034 patients receiving stented bioprosthetic valves	Early and late mortality	-	2011–2019	Early mortality was 6.1%, 1-year mortality was 11.2%, and 5-year mortality was 19.9%. Postoperative GIB was more associated with early mortality. Eight patients had GIB in the early mortality group (OR 51, 95% CI: 10.068–258.339, p < 0.001).	In-hospital GIB carried a high mortality risk in post-bioprosthetic AVR patients.
Miura et al., 2020 [31]	SAVR	Retrospective Observational	n = 201 SAVR patients	Primary: All-cause mortality and cardiac events Secondary: Adverse events, including GIB	Post-inflammatory (n = 28) vs. Congenital (n = 35) vs. Calcific/degenerative (n = 138)	2012–2015	The calcific/degenerative group, compared to the post-inflammatory group, was independently linked to a higher risk of cardiac events (HR 4.45, 95%CI 1.07–18.4, p = 0.04) and combined adverse events (HR 3.59, 95% CI 1.30–9.88, p = 0.01). Risks for combined adverse events were significantly higher in the calcific/degenerative group (HR 5.69, 95% CI 1.87–17.2, p = 0.002) and congenital group (HR 4.13, 95% CI 1.20–14.2, p = 0.02) after adjustments for age, sex, and previous procedures.	The etiology of AS could be used to predict mid-term outcomes and GIB risk after AVR.
Melgaard et al., 2021 [32]	AS	Retrospective Observational	n = 3726 AS patients with atrial fibrillation	Risk of thromboembolism and major bleeding	NOAC (n = 2357 patients) vs. warfarin (n = 1369 patients)	2013–2018	The aHR for thromboembolism and major bleeding were 1.62 (95% CI 1.08–2.45) and 0.73 (0.59–0.91) for NOAC compared with warfarin in the intention-to-treat analyses.	NOAC may be superior to warfarin in terms of bleeding profile in patients with AS and atrial fibrillation.
Van Mieghem et al., 2021 [33]	TAVI	RCT	n = 1426 severe AS patients who underwent TAVI	Composite of adverse events (all-cause mortality, myocardial infarction, ischemic stroke, systemic thromboembolism, valve thrombosis, or major bleeding)	Edoxaban (n = 713) vs. Warfarin (n = 713)	2017–2020	More patients in the edoxaban group experienced major GIB compared to the warfarin group (56 [5.4 per 100 person-years] vs. 27 [2.7 per 100 person-years]; HR, 2.03, 95% CI 1.28–3.22), despite similar rates of proton pump inhibitor use (71.7% vs. 69.0%). There was one fatal case of major gastrointestinal bleeding in the edoxaban group. Among those with major GIB, 82% (46 of 56) in the edoxaban group and 96% (26 of 27) in the warfarin group received proton pump inhibitors.	Edoxaban is non-inferior to warfarin in patients with atrial fibrillation who underwent TAVI.
Wang et al., 2023 [34]	SAVR	RCT	n = 863 pts with X-mechanical AVR	Primary: Rates of valve thrombosis or valve-related thromboembolism Secondary: Major bleeding events	Apixaban (n = 433) vs. Warfarin (n = 430)	2020–2022	The trial was stopped due to an excess of thromboembolic events in the apixaban group. A total of 26 primary end-point events occurred, 20 (in 16 participants) in the apixaban group (4.2%/patient-year, 95% CI 2.3–6.0) and 6 (in 6 participants) in the warfarin group (1.3%/patient-year, 95% CI 0.3–2.3). The difference in primary end-point rates between the apixaban and warfarin groups was 2.9 (95% CI 0.8–5.0). Major bleeding rates were 3.6%/patient-year with apixaban and 4.5%/patient-year with warfarin, with more major GI bleeding.	Apixaban is inferior to warfarin in the prevention of thromboemboli in patients with X-mechanical valves but is associated with lower major bleeding and higher GIB rates.

,

Table 2. Summary of studies evaluating GIB in mitral valve disease, listed in order of appearance in text. Abbreviations: AF, atrial fibrillation; MR, mitral regurgitation; MVR, mitral valve replacement; PH, pulmonary hypertension; PVL, paravalvular leak; TEE, transesophageal echocardiography; TMVR, transcatheter mitral valve repair; VWF, von Willebrand factor; VWFAct, von Willebrand factor activity; VWFAg, von Willebrand factor antigen.

Author, Year	Valvular Disorder	Study Type	Number of Patients	Outcomes Measured	Groups	Study Period	Notable Outcomes	Conclusions
Blackshear et al., 2014 [6]	MR	Prospective Cohort	n = 53 patients with native MR	VWFAct	Mild (n = 13), moderate (n = 14), severe (n = 26) MR	2010–2014	More VWF shearing occurred with high degrees of MR. Nine patients (17%) reported clinically significant GIB, and seven had intestinal angiodysplasia and transfusion-dependent Heyde syndrome. All measures of VWF function improved with mitral valve repair.	The risk of GIB increases with MR severity. Acquired VWF syndrome may be reversible with mitral valve surgery and lead to lower rates of GIB.
Meindl et al., 2021 [35]	MR	Prospective Observational	n = 85 patients with moderate to severe MR	VWFAct, VWFAg, and factor VIII expression before and 4 weeks after TMVR	-	2017–2020	VWFAct, VWFAct/vWFAg ratio, and VWFAg values did not change after TMVR. A significantly lower VWFAct/VWFAg ratio in primary MR versus secondary MR was observed both at baseline (p = 0.022) and 4 weeks following the TMVR procedure (p = 0.003). A mean mitral valve gradient ≥ 4 mmHg after TMVR had significantly lower VWFAct/VWFAg ratios as compared to patients with a mean mitral valve gradient. Alterations of the VWFAct/VWFAg ratio did not translate into a greater risk for bleeding events.	TMVR does not improve vWF levels or VWF activity 4 weeks after the procedure. Bleeding events in the short term after TMVR were rare despite a high perioperative risk, suggesting a superior safety profile of percutaneous MVR.
Lane et al., 2019 [36]	MR	Retrospective Observational	n = 122 pre-TMVR patients with AF	CHA2DS2-VASc and HAS-BLED risk scores	-	2014–2017	The cohort demonstrated an average CHA2DS2-VASc score of 3.99. A total of 53 and 47% of TMVR patients were at high or intermediate risk of bleeding, respectively, according to their HAS-BLED score.	TMVR candidates demonstrate an intermediate to high risk of bleeding according to their HAS-BLED score.
Tabrizi et al., 2023 [37]	MVR	Retrospective Observational	n = 442 MitraClip patients	Association between TEE procedure duration and TEE-related complications	-	2015–2022	TEE-related complications were observed in 17 patients (3.8%). Dysphagia was most common (n = 9/17, 53%), followed by new gastroesophageal reflux (n = 6/17, 35%) and odynophagia (n = 3/17, 18%). No esophageal perforations or upper gastrointestinal bleeds were appreciated. TEE procedure duration was not statistically different between the those with complication and those without.	TEE procedure duration time is not associated with GIB in patients undergoing MitraClip.
Grasso et al., 2018 [38]	MR	Prospective Observational	n = 322 MitraClip patients	Rehospitalization rates and causes	-	-	Early rehospitalization rates (<30 days) occurred in 27% of cases. Cardiovascular (66%), anemia (6%), and GIB (6%) were the most frequent causes of 1-year rehospitalization.	Significant GIB can lead to hospitalization within one year of MitraClip procedure in a small percentage of patients.
Khan et al., 2021 [39]	MR	Retrospective Cohort	n = 13,220 encounters from the National Inpatient Sample	Outcomes of TMVR	PH (n = 6610) vs. No PH (n = 6610; propensity-matched)	2014–2018	In-hospital mortality (3.6% vs. 1.9%, p < 0.01) and complications including blood transfusions (3.6% vs. 1.7%, p < 0.01), GIB (1.4% vs. 1%, p = 0.04), vascular complications (5.3% vs. 3.3%, p < 0.01), vasopressors use (2.9% vs. 1.7%, p < 0.01), and pacemaker placement (1.3% vs. 0.8%, p = 0.01) were significantly higher for encounters with PH.	Comorbid PH increases the risk of GIB in patients with MR undergoing TMVR.
Fukunaga et al., 2018 [40]	MR	Retrospective Cohort	n = 66 patients (ages 50 to 69) status post-redo mitral valve replacement	Early and late survival, freedom from thromboembolic events, and valve-related complications	Mechanical valve (n = 44) vs. Biological valve (n = 22)	1990–2015	Mechanical redo mitral valve replacement was associated with improved 5- and 10-year survival rates and freedom from thromboembolic events at 5 and 10 years. GIB was only observed in one biological valve.	The rates of GIB with mechanical or biological valves appear similar, and both could be considered in high-bleed-risk patients. However, mechanical MVR has superior survival and freedom from thromboembolic event rates.
Rokui et al., 2024 [41]	MVR	Retrospective, Multicenter, Propensity-Matched Cohort	n = 794 mitral valve replacement patients	10-year survival, freedom from intervention, and postoperative complication rate	Mechanical valve (age < 65 years: n = 226; Age 65–75 years: n = 171) vs. Bioprostethic valve (age < 65 years: n = 226; Age 65–75 years: n = 171)	2000–2017	Ten-year survival (78.2% vs. 69.8%, p = 0.029) and ten-year freedom from reintervention (96.2% vs. 81.3%, p < 0.001) were superior with mechanical valves versus bioprosthetic valves in patients < 65 years. There were no differences between mechanical and bioprosthetic valves in 10-year survival (64.6% vs. 60.8%, p = 0.86) or 10-year freedom from reintervention (94.0% vs. 97.2%, p = 0.23) in patients between 65 and 75 years. Rates of postoperative stroke, gastrointestinal bleeding, renal failure, and permanent pacemaker insertion were similar.	Mechanical and bioprosthetic mitral valve replacements demonstrate similar low rates of GIB at 10-year follow-up.
Uchino et al., 2022 [42]	MVR	Retrospective Observational	n = 583 mechanical mitral valve replacement patients	Very-long-term outcomes	-	1987–2015	Survival rates at 10, 15, 20, and 25 years were 81.42%, 69.27%, 56.34%, and 45.03%, respectively. Thromboembolism was observed in 38 patients, intracranial hemorrhage in 26, and GIB in 9. Major paravalvular leaks were observed in 32 patients. The cumulative incidence rate of major paravalvular leaks at 10, 15, 20, and 25 years was 3.7%, 5.6%, 6.4% and 10.4%, respectively.	GIB is a rare very-long-term complication associated with mechanical mitral valve replacement.
Belahnech et al., 2024 [43]	MR	Retrospective Cohort	n = 96 patients who underwent percutaneous mitral PVL closure	Recurrent GIB	Residual PVL (>mild) (n = 45) vs. No residual PVL (≤mild) (n = 51)	2010–2022	Recurrent GIB occurred in four patients with a history of GIB (4/11) and three without (3/85). All patients with recurrent GIB had residual PVL. Angiodysplasia was the most common cause of recurrent GIB (86%).	Residual PVL following MVR is associated with recurrent GIB, most commonly due to angiodysplasia.

,

Table 3. Summary of studies evaluating GIB in tricuspid valve disease, listed in order of appearance in text. Abbreviations: AF, atrial fibrillation; GDMT, goal-directed medical therapy; GIB, gastrointestinal bleeding; NOAC, non-vitamin K oral anticoagulant; RRT, renal replacement therapy; TR, tricuspid regurgitation; TTVI, transcatheter tricuspid valve implantation.

Author, Year	Valvular Disorder	Study Type	Number of Patients	Outcomes Measured	Groups	Study Period	Notable Outcomes	Conclusions
Cai et al., 2020 [5]	TR	Retrospective	n = 124 patients with symptomatic severe TR	Reasons for hospitalization, cause and time of death, GIB, stroke, myocardial infarction, and new initiation of RRT.	TTVI + GDMT (n = 53) vs. GDMT (n = 71)	2015–2019	TTVI and GDMT improved TR symptoms and functional status and were associated with lower rates of all-cause mortality, AKI, and GIB, compared to GDMT alone.	Adjunct TTVI reduces the risk of GIB in patients with severe TR.
Yang et al., 2024 [44]	TR	Retrospective	n = 1215 patients with moderate to severe TR and AF	Primary: Ischemic stroke, systemic embolic events, hospitalizations for major bleeding Secondary: Intracranial hemorrhage, gastrointestinal bleeding, all-cause mortality	NOAC (n = 724) vs. warfarin (n = 491)	2010–2020	NOACs demonstrated comparable efficacy for ischemic stroke, systemic embolic events, and major bleeding, including GIB.	NOACs are comparable to warfarin for GIB in patients with moderate to severe TR and AF.

and

Table 4. Summary of studies evaluating GIB in multiple valve disease, listed in order of appearance in text. Abbreviations: AF, atrial fibrillation; AS, aortic stenosis; AVR, aortic valve replacement; BHV, bioprosthetic heart valve; CI, confidence interval; DVR, double valve replacement; GIB, gastrointestinal bleeding; HR, hazard ratio; MI, myocardial infarction; MR, mitral regurgitation; MV, mitral valve; NOAC, non-vitamin K oral anticoagulant; RR, risk ratio; TIA, transient ischemic attack; TVP, tricuspid valvuloplasty; TVR, tricuspid valve replacement; VHD, valvular heart disease; VKA, vitamin K antagonist; VWF, von Willebrand factor; VWFAct, von Willebrand factor activity; VWFAg, von Willebrand factor antigen.

Author, Year	Valvular Disorder	Study Type	Number of Patients	Outcomes Measured	Groups	Study Period	Notable Outcomes	Conclusions
Nakatsu et al., 2015 [45]	BHV	Retrospective Observational	n = 77 dialysis patients status post-BHV	Hemorrhage frequency and survival	-	1991–2011	The overall estimated Kaplan–Meier survival after 3, 5, and 7 years was 66.6 ± 5.6, 51.1 ± 6.3, and 34.4 ± 6.8%, respectively. Seventeen (22%) bleeding events were observed (5.9% per patient-year) in the follow-up period. Six of the nine patients with cerebral hemorrhages and two of the six with gastrointestinal hemorrhages died. There were no differences in the frequencies of hemorrhage between the patients treated with bioprostheses and mechanical valves.	GIB is a notable cause of hemorrhage and mortality in dialysis patients status post-BHV.
Blackshear et al., 2016 [46]	AS/MR	Prospective Cohort	n = 136 patients with varying valvular disease	Severity of VWF dysfunction	Normally functioning surgical or transcatheter aortic valve (n = 26) vs. Dysfunctional aortic valve replacement (n = 24) vs. Normally functioning mitral valve replacement or repair (n = 36) vs. Dysfunctional mitral valve replacement or repair (n = 19) vs. Native AR without stenosis (n = 31)	2010–2015	Normal AVR showed higher VWFAct-to-VWFAg ratio compared to dysfunctional aortic valve replacement (p < 0.001), as did normal mitral valve replacement or repair compared to dysfunctional mitral valve replacement or repair (p = 0.005). GIB was noted in 25% of patients with aortic prosthesis dysfunction and 26% with mitral prosthesis/repair dysfunction and was associated with a lower normalized VWF multimer ratio than in patients without bleeding. Gastrointestinal angiodysplasia was noted in 83% of bleeding patients with dysfunctional aortic prostheses and in 60% of bleeding patients with dysfunctional mitral prostheses/repair.	Acquired abnormalities of VWF multimers are associated with aortic and mitral prosthesis dysfunction, with occasional GIB and gastrointestinal angiodysplasia.
Briasoulis et al., 2018 [47]	Native VHD	Retrospective Cohort	n = 103,733 propensity-matched patients with new AF	All-cause mortality, ischemic stroke, major bleeding, and MI	Nonvalvular AF (n = 85,596) - Dabigatran (n = 13,522) - Rivaroxaban (n = 14,257) - Warfarin (n = 57,817) Valvular AF (n = 18,137) - Dabigatran (n = 1979) - Rivaroxaban (n = 2027) - Warfarin (n = 14,131)	2011–2013	Both dabigatran and rivaroxaban were associated with significantly lower risk of death in patients with VHD with AF (dabigatran versus warfarin: HR 0.71; 95% CI 0.52–0.98; p = 0.038; rivaroxaban versus warfarin: HR 0.68; 95% CI 0.49–0.95; p = 0.022). Non-GIB was significantly reduced with dabigatran and rivaroxaban versus warfarin in those with VHD (dabigatran versus warfarin: HR 0.17; 95% CI 0.06–0.49; p = 0.001; rivaroxaban versus warfarin: HR 0.37; 95% CI 0.17–0.84; p = 0.017). Ischemic stroke and GIB rates did not differ between rivaroxaban, dabigatran, and warfarin in patients with VHD. The effects of the three anticoagulants on outcomes were comparable in patients with and without VHD and AF.	NOACs can safely be used in high-bleeding risk AF patients with or without VHD.
Duan et al., 2021 [48]	BHV	Retrospective Cohort	n = 2672 AF patients with BHV	Ischemic stroke, systemic embolism, TIA, and adverse effects	NOAC (n = 439) vs. warfarin (n = 2233)	2011–2020	No significant difference in ischemic stroke, embolism, or TIA between NOACs and warfarin (HR 1.19, 95% CI 0.96–1.48, p = 0.11). NOACs were associated with a lower risk of the intracranial hemorrhage, GIB, and other bleeds (HR 0.69, 95% CI 0.56–0.85, p < 0.001).	NOACs are non-inferior to warfarin in thromboembolic prevention in patients with AF and BHV but should be preferred given their more favorable adverse bleeding effect profile, especially in high-bleeding-risk patients.
Suppah et al., 2023 [49]	BHV	Systematic Review with Meta-Analysis	n = 74,660 patients with BHV and AF from 28 studies (I2 = 0%, p < 0.10)	Primary: all-cause bleeding, major bleeding, stroke/systemic embolization, all-cause mortality Secondary: GIB, thromboembolic events, intracranial bleeding	NOACs (n = variable) vs. VKA (n = variable)	2016–2022	NOACs decrease the risk of all-cause bleeding (RR 0.80, 95% CI 0.75–0.85, p > 0.00001), stroke and systemic embolization (RR 0.89, 95% CI 0.80–0.99, p = 0.03), and intracranial bleeding outcomes (RR 0.62, 95% CI 0.45–0.86, p = 0.004) compared with VKA. There was no significant difference in major bleeding (RR = 0.92, 95% CI 0.84 1.02, p = 0.10) and all-cause mortality outcomes (RR = 0.96, 95% CI 0.85–1.07, p = 0.43), respectively. No significant difference was observed in gastrointestinal bleeding.	NOACs are comparable to VKAs for GIB in patients with a BHV and AF.
Liu et al., 2015 [50]	Valve Surgery	Retrospective Cohort	n = 1240 patients who received valve surgeries (MV replacement, mitral valvuloplasty, TVR, AVR, DVR, TVP)	Success and adverse effect rates	Age ≤ 60 (n = 603) (BHV, n = 17) vs. Age > 60 (n = 637) BHV, n = 80)	2004–2014	The age > 60 group had higher application rates of bioprosthetic valves, times of auxiliary ventilation, and hospitalization stay lengths (p < 0.05). No difference in low mortality rates (2.7% vs. 3.1%, p > 0.05) was appreciated. No difference in the incidence or mortality of upper GIB between groups (number of cases, 10/603 vs. 14/637; mortality, 10% vs. 7.1%; p = 0.456).	The age-related incidence and mortality of GIB in all patients undergoing valvuloplasty or valve replacement are similar.

.

3.1. Gastrointestinal Bleeding and Aortic Valve Dysfunction

The association between the risk of GIB and aortic stenosis (AS) has been well documented and reported in the literature. A retrospective observational study by Aldiabat et al. in 2024 identified 85,090 patients with gastric antral vascular ectasia using the National Inpatient Sample from 2016 to 2019 and found those with aortic stenosis to have a twofold increased risk of the disease (adjusted odds ratio (aOR) 2.08, 95% CI 1.94–2.22, p < 0.001) [8]. A Taiwanese multi-center retrospective study found AS associated with persistent or recurrent GIB of unknown origin after negative endoscopic results in 765 patients (p = 0.047) after adjusting for antithrombotic agents [9]. In another 2021 retrospective study by Waldschmidt et al., 190 of 2548 transcatheter aortic valve intervention (TAVI) patients (7.8%) had a history of GIB, with roughly 2% attributed to Heyde syndrome. Patients diagnosed with Heyde syndrome required more blood transfusions (50.0% vs. 31.9%, p = 0.03), and the rate of recurrent GIB was higher compared to those without (39.8% vs. 21.2%, p = 0.03) [10].

AS has been strongly correlated with gastrointestinal angiodysplasia. In 2017, a retrospective study of 32,079 patients by Jehangir et al. found intestinal angiodysplasia to be associated with AS in approximately 7% of cases. This association remained statistically significant when adjusted for age and known risk factors. The unadjusted odds of aortic valve disease in association with bleeding intestinal angioectasia were 4.95 (95% CI 4.43–5.54, p < 0.001), compared to those without bleeding angioectasia [11]. Additionally, Tamura et al. prospectively enrolled 31 patients with severe AS and found a loss of the VWF multimer in two-thirds of patients. The multimer ratio was inversely correlated with the peak aortic gradient (R = −0.58, p = 0.0007). Among patients with anemia (12), four were found to have gastrointestinal angiodysplasia, resulting in a diagnosis of Heyde syndrome [12].

Nineteen studies evaluated surgical and transcatheter valvular interventions for AS in the setting of Heyde syndrome. In 2020, Selvam et al. prospectively enrolled 29 patients with AS and analyzed VWF plasma antigen and activity before and after aortic valve replacement (AVR). The mean VWF antigen significantly increased 3 to 5 days and 6 months following surgery (2 ± 1.76 IU/mL vs. 4.37 ± 2.61 IU/mL vs. 3.23 ± 2.49 IU/mL, respectively; p < 0.05), demonstrating the lasting benefits of AVR [13]. Moreover, a 2014 systematic review with meta-analysis by Jackson et al. found that AVR (pooled OR 0.19, 95% CI 0.11–0.30) had a lower risk of rebleeding at four years compared to somatostatin analogs (pooled OR 14.5, 95% CI 5.9–36) or hormonal therapy (pooled OR 1.0, 95% CI 0.5–1.96) in a total of 623 patients with gastrointestinal angiodysplasia lesions [14].

Another 2015 study by Caspar et al. prospectively enrolled 49 patients with severe AS referred for TAVI and measured VWF levels at baseline and 1 week after the procedure. The mean aortic transvalvular gradient demonstrated an inverse relationship with VWF antigen levels (R = −0.29; p < 0.05), VWF ristocetin cofactor activity (R = −0.402; p = 0.006), and VWF collagen-binding activity (R = −0.441; p = 0.005). One week after the procedure, a significant increase in VWF antigen (3.32 vs. 2.29 IU/mL, p < 0.001), VWF ristocetin cofactor activity (2.98 vs. 1.86 IU/mL, p < 0.001), and VWF collagen-binding activity (3.16 vs. 2.16 IU/mL, p < 0.001) was demonstrated. These findings support the evidence that hemostasis parameters improve after TAVI [15].

In another study from 2022, Goltstein et al. performed a retrospective observational study that included 70 patients with Heyde syndrome who underwent TAVI between 2008 and 2020. The incidence of GIB (62% of patients, 95% CI 50–74%) and bleeding episodes [3.2 (95% CI 2.5–4.2) vs. 1.6 (95% CI 1.2–2.2), p = 0.001] decreased following TAVI. Mean hemoglobin levels also increased from 10.3 (95% CI, 10.0–10.8) to 11.3 (95% CI, 10.8–11.6) g/dL (p = 0.007). Between 1 and 5 years after TAVI, 83% of patients no longer experienced gastrointestinal bleeding [16].

A 2017 retrospective cohort study by Tjahjadi et al. identified 9562 patients who underwent coronary artery bypass grafting (CABG) or AVR. The study found 18.4% of all patients to be anemic (95% CI 17.7–19.2%), including 30.1% in the AVR group (95% CI 27.9–32.5%) and 16.2% (95% CI 15.4–17.0%) in the CABG group. The study concluded that anemia may be more prevalent in patients with AS than those who underwent CABG, possibly due to chronic GIB from Heyde syndrome [17].

Sugino et al. released a multicenter retrospective study in 2023 that reviewed the endoscopic features of angiodysplasia in patients with severe AS. In this study of 325 patients who underwent TAVI, patients with moderate to severe anemia demonstrated significantly more angiodysplasia (38.3% vs. 7.7%; p < 0.0001) and active bleeding (23.4% vs. 0%; p < 0.01) [18]. Another retrospective study of 841 TAVI patients from a Canadian tertiary care center by Stanger et al. analyzed the risk of upper GIB. The incidence of upper GIB was 2%, with a nonsignificant trend towards increasing age, as those with upper GIB were an average of 2.8 years older than patients without UGIB. Additionally, the risk of upper GIB in patients receiving triple antithrombotic therapy was 10-fold higher than those not on it (11.8% vs. 1%). These studies show that patients with severe AS who underwent TAVI and a history of anemia or upper GIB continue to have some increased risk of GIB [19].

A 2023 retrospective propensity-matched national analysis of 320,353 TAVI patients by Manasrah et al. had their readmissions and outcomes compared. A total of 6193 of these patients were readmitted for GIB between 2016 and 2020, with higher readmission rates at 30, 90, and 180 days observed in patients with GIB (aOR 6.35, 95% CI 5.37–7.52, p < 0.0001) [20]. In another 2022 retrospective study, Zahid et al. identified 216,023 TAVI patients and found that 2188 (1%) patients were readmitted for GIB between 2011 and 2018. Histories of peptic ulcer disease (OR 8.74, 95% CI 6.69–11.43), colorectal cancer (OR 7.89, 95% CI 5.33–11.70), chronic kidney disease (OR 1.27, 95% CI 1.14–1.41), liver disease (OR 1.83, 95% CI 1.53–2.19), and atrial fibrillation (OR 1.63, 95% CI 1.49–1.78) were associated with readmission. Patients with GIB demonstrated a higher mortality rate (12.1% vs. 3.2%, p < 0.01) than those without. The study concluded that hospitalization of TAVI patients for GIB can be associated with higher readmission and mortality rates compared to those without GIB [21]. In 2014, Laflamme et al. retrospectively analyzed the left axillary artery approach as opposed to the femoral or transapical approach in 18 TAVI patients. The axillary approach had a higher procedural success and lower rates of complications. Only one case of upper GIB was appreciated [22].

A retrospective observational study by Brown et al. in 2022 identified 1192 TAVI patients with abnormal hemoglobin and hematocrit levels. A total of 13.8% had severe GIB, with higher Society of Thoracic Surgeons-Predicted Risk of Mortality scores (8.8 ± 5.3 vs. 7.6 ± 4.5, p = 0.002) compared to those without GIB. A total of 79.3% had resolution of GIB after TAVI. Those with recurrent GIB had higher aortic valve mean pressure gradients after TAVI (15.0 ± 5.3 vs. 9.0 ± 4.3), suggesting that higher trans-aortic valve pressure gradients following TAVI may predispose patients to recurrent bleeding events [23].

In 2020, Yerasi et al. retrospectively analyzed the rate of next-day discharges in 49,742 TAVI patients between 2012 and 2016. The rate of next-day discharges post-TAVI has increased from 1.5% in 2012 to 12.2% in 2016 with similar readmission rates; however, the rate of GIB has increased since 2016. A significant predictor of readmission was anemia (OR 1.30, 95% CI 1.10–1.54, p = 0.002), suggesting that post-TAVI patients with anemia may need to be monitored in the hospital for more than one day after the intervention [24].

Another retrospective study by Iyengar et al. in 2018 compared hospitalization rates and reasons in 43,357 TAVI patients to 310,013 surgical aortic valve replacement (SAVR) patients. A total of 3.3% of TAVI patients were re-hospitalized for GIB, compared to 1.5% SAVR patients (p < 0.001), which remained statistically significant after multivariable adjustment (aOR 1.54, 95% CI 1.38–1.71, p < 0.001). This study suggests that TAVI patients may have more frequent readmissions for late GIB than SAVR [25]. In 2019, Desai et al. performed a retrospective, propensity-matched analysis of 130,487 patients’ status post-TAVI or SAVR. The investigators observed no significant differences between the SAVR and TAVI groups regarding all-cause mortality (11.7% vs. 10.0%, p = 0.51) and hospital charges (USD 400,816 vs. USD 381,065, p = 0.48). TAVI patients exhibited significantly lower rates of stroke (1.7% vs. 10.0%, p < 0.001), periprocedural myocardial infarction (3.3% vs. 6.7%, p = 0.06), and blood transfusion (1.7% vs. 11.7%, p < 0.001) compared to SAVR patients [26].

A 2023 systematic review with meta-analysis by Goltstein et al. compared the effectiveness of TAVI versus SAVR in 1054 patients with Heyde syndrome and assessed their status post-TAVI or SAVR. AVR, overall, led to the recovery of VWF deficiency in 85.9% (95% CI 79.1–90.7%, I² = 26%) of patients within 24 h post-procedure; in 89.5% (95% CI 74.1–96.2%, I² = 80%) between 24 and 72 h; in 92.2% (95% CI 84.0–96.3%, I² = 58%) between 3 and 21 days; and in 87.4% (95% CI 67.2–95.9%, I² = 84%) between 4 weeks and 2 years. Subgroup analysis further showed that GIB cessation was achieved in 82.0% (95% CI 71.8–89.1%, I² = 0%) after SAVR and 64.2% (49.8–76.5%, I² = 69%) after TAVI (p = 0.003). Additionally, AVR was linked to the complete cessation of GIB in 72.8% (95% CI 62.2–81.3%, I² = 59%) of patients, with follow-up periods ranging from a median of 12 to 108 months [27].

In 2023, Matsuura et al. compared bleeding rates in 837 patients who received a bioprosthetic valve replacement to 692 who received a mechanical valve. GIB occurred more frequently in the mechanical valve group as compared to the bioprosthetic valves (number of events: 86, hazard ratio (HR) 2.071, 95% CI 1.243–3.451, p = 0.0005) [28]. Additionally, Kyto et al. in 2019 also retrospectively reported bleeding rates in 296 patients who received a mechanical valve and 3931 who received a bioprosthetic valve. The 10-year major bleeding rates were 37.0% with mechanical valves and 18.8% with biologic valves (HR 1.77, 95% CI 1.25–2.49; p = 0.001, number needed to harm = 7.4). GIB was significantly more frequent with mechanical valve prosthesis (26.5% vs. 8.9%, HR 2.63; 95% CI 1.63–4.23; p < 0.001) [29]. Another 2023 retrospective study by Rashvand et al. found GIB to be related to early mortality in 1034 patients who received stented bioprosthetic valves (OR 51, 95% CI 10.068–258.339, p < 0.001) [30].

In 2020, Miura et al. retrospectively evaluated the outcomes of SAVR by etiology of AS in 201 patients. Patients with AS due to calcific causes or degeneration demonstrated a higher risk of cardiac events (HR 4.45, 95%CI 1.07–18.4; p = 0.04) and combined adverse events, including GIB (HR 3.59, 95% CI 1.30–9.88, p = 0.01), compared to the post-inflammatory group. Risks for combined adverse events were significantly higher in the calcific/degenerative group (HR 5.69, 95% CI, 1.87–17.2; p = 0.002) and congenital group (HR 4.13, 95% CI 1.20–14.2, p = 0.02) after adjusting for age, sex, and previous procedures [31].

Three studies investigated the choice of anticoagulation in patients with valvular abnormalities and aortic valve disease. A 2021 retrospective study by Melgaard et al. compared 2357 patients with AS and atrial fibrillation receiving non-vitamin K oral antagonists to 1369 patients receiving warfarin. GIB occurred less frequently in the NOAC group than warfarin (aHR 0.73, 95% CI 0.59–0.91) [32]. A 2021 multicenter randomized controlled trial by Van Mieghem et al. evaluated edoxaban versus warfarin in 713 and 713 TAVI patients with atrial fibrillation, respectively. GIB was more prevalent in those receiving edoxaban compared to warfarin (56 [5.4 per 100 person-years] vs. 27 [2.7 per 100 person-years]; HR, 2.03, 95% CI 1.28–3.22), despite similar rates of proton pump inhibitor use (71.7% vs. 69.0%). There was one fatal case of major GIB in the edoxaban group [33].

Finally, a 2023 prospective study by Wang et al. analyzed the bleeding rate in 433 patients with an X-mechanical AVR on apixaban and 430 on warfarin. A total of 17 major bleeding events occurred in 11 participants in the apixaban group (3.6%/patient-year) and 21 events in 18 participants in the warfarin group (4.5%/patient-year) (HR 0.6, 95% CI 0.3–1.3). In this study, Apixaban did not demonstrate noninferiority to warfarin [34]. Therefore, the choice of AC should be individualized and considered on a case-by-case basis.

3.2. Gastrointestinal Bleeding and Mitral Valve Dysfunction

Recent studies have correlated mitral regurgitation (MR) and mitral valve repair (MVR) with GIB. One prospective observational study by Blackshear et al. in 2014 correlated the severity of MR to the risk of GIB. The investigators aimed to assess the prevalence and severity of acquired von Willebrand syndrome in 13 patients with mild, 14 with moderate, and 26 with severe MR. Loss of the highest-molecular-weight von Willebrand factor (VWF) multimers occurred in 8%, 64%, and 85%, respectively, to mild, moderate, and severe MR. The ratios of VWF latex activity to antigen were 0.92 (IQR 0.83–0.97), 0.85 (IQR 0.76–0.89), and 0.79 (IQR 0.75–0.82), respectively (all p < 0.001). Nine patients reported clinically significant bleeding and seven had intestinal angiodysplasia and transfusion-dependent GIB (Heyde syndrome), with a median of 20 (IQR 10–33; range 4–50) transfusions needed. All measures of VWF function reported above improved significantly in patients who underwent mitral valve repair (n = 13) or replacement (n = 7) [6]. This study demonstrates the impact of MR severity on the risk of GIB and proposes that mitral valve surgery could reverse any acquired von Willebrand syndrome associated with MR.

In 2021, Meindl et al. prospectively enrolled 123 patients with moderate to severe MR and confirmed the findings of Blackshear et al. and further compared primary versus secondary etiologies of MR. Patients with primary MR had significantly lower VWF activity (VWFAct)/VWF antigen (VWFAg) ratios as compared to patients with secondary MR at baseline (p = 0.022) and at four weeks’ follow-up (p = 0.003). Also, patients with a postprocedural mean mitral valve gradient ≥ 4 mmHg had significantly lower VWFAct/VWFAg ratios than patients with a mean mitral valve gradient < 4 mmHg (p = 0.001). Only two in-hospital GIBs (2.4%) and one GIB at 4-week follow-up after MitraClip (1.2%) were reported [35]. This study suggests primary MR may correlate more strongly with acquired von Willebrand syndrome than secondary MR; however, the clinical translation to significant GIB may not be significant.

Four studies examined the relationship between GIB rates and MVR. In 2019, Lane et al. identified 122 patients with atrial fibrillation undergoing transcatheter MVR to assess their bleeding risk and determine whether they could benefit from pre-MVR implantation of an alternative ischemic stroke prophylaxis like a left atrial appendage closure device. According to their HAS-BLED score, 53 and 47% of transcatheter MVR patients were at high or intermediate risk of bleeding. An average CHA2DS2-VASc score of 3.99 was appreciated. Based on risk assessments, the investigators concluded that 50% of patients undergoing transcatheter MVR would also qualify for a left atrial appendage closure device. Given the need for systemic anticoagulation in patients with atrial fibrillation undergoing MVR, risk assessments and bleeding mitigation strategies could be implemented before MVR [36].

Another retrospective study by Tabrizi et al. in 2023 investigated the association between transesophageal echocardiography (TTE) procedural duration time and TTE-related complications in 442 patients who underwent mitral valve transcatheter edge-to-edge repair. A total of 17 complications occurred (3.8%), with the most common being dysphagia (53%), new gastroesophageal reflux (35%), and odynophagia (18%). No esophageal perforations or GIB occurred. The duration of TEE was not statistically different between those with and without complications [37]. This study highlights the safety of TEE for MVR concerning GIB.

In 2018, Grasso et al. conducted a prospective observational study with multivariate analyses to identify predictors of early (30-day) and late (30-day to 12-month) rehospitalizations in 322 MitraClip patients. Eighty-nine patients (27.6%) were readmitted to the hospital during the study period, and early rehospitalization occurred in 27%. The median time from MitraClip to rehospitalization was 99 days. Cardiovascular causes (66.3%) were the most common reasons for rehospitalizations, followed by chronic anemia (6%) and gastrointestinal bleeding (6%). Significant GIB can lead to hospitalization within one year of MitraClip in a small percentage of patients [38].

Another retrospective study by Khan et al. in 2021 compared the in-hospital outcomes of 6610 patients with and 6610 without pulmonary hypertension who underwent transcatheter MVR using data from the National Inpatient Sample. In-hospital mortality (3.6% vs. 1.9%, p < 0.01) and complications including blood transfusions (3.6% vs. 1.7%, p < 0.01), GIB (1.4% vs. 1%, p = 0.04), vascular complications (5.3% vs. 3.3%, p < 0.01), vasopressors use (2.9% vs. 1.7%, p < 0.01), and pacemaker placement (1.3% vs. 0.8%, p = 0.01) were significantly higher for encounters with pulmonary hypertension [39]. This study illustrates how comorbid pulmonary hypertension increases the risk of GIB in patients with MR undergoing transcatheter MVR.

Three studies evaluated the impact of mechanical and bioprosthetic valve types in MVR or mitral valve replacement. In 2018, Fukunaga et al. retrospectively identified 66 patients between 50 and 69 years of age who underwent a redo mitral valve replacement. Forty-four patients received a mechanical valve, while twenty-two received a bioprosthetic valve. Mechanical valves were associated with improved 5- and 10-year survival rates and freedom from thromboembolic events at 5 and 10 years. GIB was only observed in one patient who received a bioprosthetic valve. As the rates of GIB with mechanical or biological valves appear similar, both could be considered in high-bleed-risk patients. However, mechanical MVR has superior survival and freedom from thromboembolic event rates, suggesting mechanical valves should be preferred [40].

In 2024, Rokui et al. conducted a retrospective, multicenter, propensity-matched cohort study investigating the 10-year survival, freedom from intervention, and postoperative complication rates in 794 patients. Ten-year survival (78.2% vs. 69.8%, p = 0.029) and ten-year freedom from reintervention (96.2% vs. 81.3%, p < 0.001) were superior with mechanical valves compared to bioprosthetic valves in patients < 65 years old. There were no significant differences in 10-year survival rates (64.6% for mechanical valves vs. 60.8% for bioprostethic valves, p = 0.86) or in 10-year freedom from reintervention (94.0% for mechanical vs. 97.2% for bioprosthetic, p = 0.23) among patients aged 65 and 75 years. Rates of postoperative stroke, gastrointestinal bleeding, renal failure, and permanent pacemaker insertion were similar [41]. This study indicates mechanical and bioprosthetic mitral valve replacements demonstrate similarly low rates of GIB at 10-year follow-up.

Another retrospective observational study by Uchino et al. in 2022 aimed to examine the very-long-term outcomes of mechanical mitral valve replacements in 583 patients. The survival rates at 10, 15, 20, and 25 years were 81.42%, 69.27%, 56.34% and 45.03%, respectively. Thromboembolism was observed in 38 patients, intracranial hemorrhage in 26, and GIB in 9. Major paravalvular leaks were observed in 32 patients. The cumulative incidence rate of major paravalvular leaks at 10, 15, 20, and 25 years was 3.7%, 5.6%, 6.4% and 10.4%, respectively [42]. The findings from this study indicate that GIB is a rare, very long-term complication associated with mechanical mitral valve replacement.

One retrospective study by Belahnech et al. from 2024 evaluated the effect of paravalvular leaks (PVL) on recurrent GIB in 96 patients who underwent percutaneous mitral PVL closure. All closures were performed on surgically placed prosthetic mitral valves, of which 83.3% were mechanical. Patients were divided into groups based on residual (>mild) (n = 45) or no residual (≤mild) (n = 51) PVL. Recurrent GIB occurred in four patients with a history of GIB (4/11) and three without (3/85). All patients with recurrent GIB had residual PVL. Angiodysplasia was the most common cause of recurrent GIB (86%) [43]. The study’s findings show that residual PVL following MVR is associated with recurrent GIB, most commonly due to angiodysplasia.

3.3. Gastrointestinal Bleeding and Triscuspid Valve Dysfunction

A retrospective observational study by Cai et al. in 2020 aimed to identify the clinical characteristics and examine outcomes of 53 patients with symptomatic severe tricuspid regurgitation who received transcatheter tricuspid valve intervention compared to 71 patients with symptomatic severe tricuspid regurgitation who were treated with guideline-directed medical therapy. Transcatheter tricuspid valve intervention significantly improved NYHA functional class and 6 min walk distance (p < 0.001). Patients treated with goal-directed medical therapy demonstrated lower survival (46.9% vs. 75.1%, p = 0.047) and lower freedom from a heart failure hospitalization and mortality (33.2% vs. 62.7%, p = 0.027), as well as higher incidences per 100 person-year of gastrointestinal bleeding [15.58 (95% CI 8.90–25.31) vs. 4.24 (95% CI 0.85–12.37), p = 0.04] and acute kidney injury [36.98 (95% CI 26.17–50.76) vs. 14.12 (95% CI 6.76–25.96), p = 0.0001], compared with TTVI patients [5]. The study indicates that mechanical interventions for tricuspid regurgitation may reduce the incidence of GIB compared to medical management.

Another study by Yang et al. in 2024 aimed to compare the efficacy and safety of non-vitamin K oral anticoagulants (NOAC) in 724 patients with atrial fibrillation and moderate to severe tricuspid regurgitation compared to 491 treated with warfarin. NOACs were similar to warfarin for ischemic stroke and systemic embolic events (adjusted hazard ratio (aHR) 0.95, 95% CI 0.67–1.36, p = 0.79) and major bleeding (aHR 0.78, 95% CI 0.57–1.06, p = 0.11). NOACs demonstrated a lower risk of intracranial hemorrhage (aHR 0.27, 95% CI 0.14–0.53, p < 0.001) and a comparable risk for GIB (aHR 1.14, 95% CI 0.77–1.68, p = 0.52) and all-cause mortality (aHR 0.38, 95% CI 0.64–1.19, p = 0.38) [44]. The study’s findings note that medical thromboembolic prevention with DOACs is equivocal to warfarin in terms of the risk for GIB.

3.4. Other Gastrointestinal Bleeding in Valvular Diseases

Six other studies reviewed GIB in multiple valvular diseases (Table 4). In 2015, Nakatasu et al. retrospectively assessed hemorrhage frequency and survival rates in 77 dialysis patients who underwent bioprosthetic heart valve replacement. The estimated Kaplan–Meier survival at 3, 5, and 7 years was 66.6 ± 5.6, 51.1 ± 6.3, and 34.4 ± 6.8%, respectively. During the follow-up period, 17 (22%) bleeding events occurred, representing 5.9% per patient-year. Among the patients, six out of nine with cerebral hemorrhages and two out of six with gastrointestinal hemorrhages did not survive. No significant differences in the frequencies of hemorrhage between the patients treated with bioprostheses and mechanical valves were appreciated [45]. The study highlights GIB as a notable cause of hemorrhage and mortality in dialysis patients’ status post-bioprosthetic heart valve replacement.

A prospective cohort study by Blackshear et al. in 2016 aimed to determine the association between dysfunctional valve prostheses and VWF abnormalities compared with normally functioning valve prostheses, identify the severity of the VWF abnormality relative to other conditions, and describe associated bleeding and the occurrence of gastrointestinal angiodysplasia in 136 patients with various valvular abnormalities. The cohorts comprised 26 patients with a normally functioning surgical or transcatheter aortic valve, 24 with a dysfunctional aortic valve replacement, 36 with a normally functioning mitral valve replacement or repair, 19 with a dysfunctional mitral valve replacement or repair, and 31 with a native aortic regurgitation without stenosis. Normal aortic valve replacement showed a higher VWFAct-to-VWFAg ratio compared to dysfunctional aortic valve replacement (p < 0.001), as did normal mitral valve replacement or repair compared to dysfunctional mitral valve replacement or repair (p = 0.005). GIB was noted in 25% of patients with aortic prosthesis dysfunction and 26% with mitral prosthesis/repair dysfunction. It was associated with a lower normalized VWF multimer ratio compared to patients without bleeding. Gastrointestinal angiodysplasia was observed in 83% of bleeding patients with dysfunctional aortic prostheses and in 60% of bleeding patients with dysfunctional mitral prostheses or repairs [46]. The findings of the study indicate that acquired abnormalities of VWF multimers are associated with aortic and mitral prosthesis dysfunction, with occasional GIB and gastrointestinal angiodysplasia.

Three of the six studies compared NOACs to vitamin K antagonists (VKAs) in the co-management of valvular heart disease and atrial fibrillation. In 2018, Briasoulis et al. retrospectively compared 85,596 patients with nonvalvular atrial fibrillation to 18,137 with native valvular atrial fibrillation. The investigators found both dabigatran and rivaroxaban to be associated with a significantly lower risk of death in patients with valvular heart disease with atrial fibrillation (dabigatran versus warfarin: HR 0.71; 95% CI 0.52–0.98; p = 0.038; rivaroxaban versus warfarin: HR 0.68; 95% CI 0.49–0.95; p = 0.022). Non-GIB rates were significantly lower with dabigatran and rivaroxaban compared to warfarin in patients with valvular heart disease (dabigatran versus warfarin: HR 0.17; 95% CI 0.06–0.49; p = 0.001; rivaroxaban versus warfarin: HR 0.37; 95% CI 0.17–0.84; p = 0.017). However, the rates of ischemic stroke and GIB rates did not differ among rivaroxaban, dabigatran, and warfarin in this population. The effects of these three anticoagulants on outcomes were similar for patients both with and without valvular heart disease and atrial fibrillation [47]. The results of this study suggest that NOACs can safely be used in high-bleeding risk atrial fibrillation patients with or without valvular heart disease.

In 2021, Duan et al. also retrospectively compared NOACs to VKAs but did so in 2672 patients with atrial fibrillation and a bioprosthetic heart valve. The investigators found no significant difference in ischemic stroke, embolism, or TIA between NOACs and warfarin (HR 1.19, 95% CI 0.96–1.48, p = 0.11). NOACs were associated with a lower risk of intracranial hemorrhage, GIB, and other bleeding (HR 0.69, 95% CI 0.56–0.85, p < 0.001) [48]. The findings suggest NOACs are non-inferior to warfarin in thromboembolic prevention in patients with atrial fibrillation and bioprosthetic heart valves but should be preferred given their more favorable adverse bleeding effect profile, especially in high-bleeding-risk patients.

A systematic review with meta-analysis by Suppah et al. conducted in 2023 compared the efficacy and adverse effect profiles of NOACs and VKAs of 74,660 patients with BHV and AF from 28 studies (I² = 0%, p < 0.10). NOACs were found to decrease the risk of all-cause bleeding (RR 0.80, 95% CI 0.75–0.85, p > 0.00001), stroke and systemic embolization (RR 0.89, 95% CI 0.80–0.99, p = 0.03), and intracranial bleeding outcomes (RR 0.62, 95% CI 0.45–0.86, p = 0.004) compared with VKA. No significant differences in major bleeding (RR = 0.92, 95% CI 0.84 1.02, p = 0.10), GIB, or all-cause mortality outcomes (RR = 0.96, 95% CI 0.85–1.07, p = 0.43) were appreciated [49]. These findings suggest that NOACs are comparable to VKAs concerning GIB in patients with a bioprosthetic heart valve and atrial fibrillation.

Finally, one retrospective cohort study by Liu et al. conducted in 2015 studied the success and adverse effect rates of valvuloplasties and valve surgeries in 1240 patients. Patients who received mitral valve replacements or valvuloplasty, tricuspid valve replacement or valvuloplasty, aortic valve replacement, or double valve replacement were included and divided into age < 60 years or age > 60 years at the time of surgery. The age > 60 group had higher application rates of bioprosthetic valves, times of auxiliary ventilation, and hospitalization stay lengths (p < 0.05). No difference in low mortality rates (2.7% vs. 3.1%, p > 0.05) was appreciated. There was also no difference in the incidence or mortality of upper GIB between groups (number of cases, 10/603 vs. 14/637; mortality, 10% vs. 7.1%; p = 0.456) [50]. The findings of this study suggest that the age-related incidence and mortality of GIB in all patients undergoing valvuloplasty or valve replacement are similar.

4. Discussion and Narrative Synthesis

4.1. Exploring the Association Between Valvular Disease and Gastrointestinal Bleeding

Gastrointestinal bleeding related to valvular heart disease has been well explored in the literature, especially in the context of angiodysplasia related to aortic stenosis or Heyde syndrome, as found by Jehangir et al. [11]. Although classically considered a rare disease, with a prevalence of 1.8% as identified by Waldschmidt et al., studies by Tjahjadi et al. and Tamura et al. suggest Heyde syndrome may be underdiagnosed, given the high rates of chronic anemia in patients with AS undergoing TAVI [10,12,17]. Blackshear et al. also concluded that dysfunctional aortic prosthesis to be associated with angiodysplasia in approximately 83% of patients with GIB [46]. Aside from angiodysplasia, Aldiabat et al. also suggested that patients with aortic stenosis have a two-fold increased risk of gastric antral vascular ectasia compared to those without aortic stenosis [8].

Other valvular diseases have been implicated in the development of Heyde syndrome. Blackshear et al. concluded that the high-shear environment of moderate to severe mitral regurgitation is sufficient enough to produce prevalent perturbations of VWF activity and, therefore, GIB; however, Meindl et al. found these alterations in the ratio of VWF activity to VWF antigen to not translate into a greater risk for bleeding events [6,35]. In addtion, Cai et al. found valvular interventions for tricuspid regurgitation to reduce the risk of GIB in patients with severe TR [5]. However, none of the available studies investigated the etiologies of GIB or the presence of acquired von Willebrand syndrome in patients with tricuspid regurgitation.

4.2. Pathophysiology of Gastrointestinal Bleeding in Valvular Disease

The pathophysiology of GIB in valvular heart disease is well understood and well described in the literature. It has been best described in the setting of calcific aortic stenosis, largely due to the prevalence of the disease; however, mitral and tricuspid regurgitation have also been implicated (Figure 2) [6]. The triad of GIB, aortic stenosis, and acquired von Willebrand syndrome, also known as Heyde syndrome, is strongly associated with coagulopathies, angiodysplasia, and subsequent bleeding. Bleeding in Heyde syndrome is thought to be multifactorial, with acquired von Willebrand disease playing a pivotal role. Von Willebrand factor (VWF) is a large multimeric glycoprotein synthesized by endothelial cells and megakaryocytes. It is cleaved into low- and high-molecular-weight polymers. The large multimers facilitate optimal platelet adhesion and aggregation in hemostasis. High shear stress from a calcified aortic valve promotes fragmentation of large von Willebrand factor polymers, leading to uncoiling and increased proteolytic cleavage by ADAMSTS13, resulting in the loss of high-molecular-weight VWF, reduced platelet adhesion, and more potential for GIB [7,51].

Angiodysplasia is a vascular lesion in the mucosal or submucosal layers caused by chronic ischemia, hypoxia, or vascular obstruction. The development of angiodysplasia is driven by increased angiogenesis due to the upregulation of vascular endothelial growth factor (VEGF). VWF, released by endothelial cells, can modulate VEGF receptor signaling via integrin αvβ3, which both promotes and represses angiogenesis. VWF stabilizes integrin αvβ3, reducing VEGF receptor signaling and activity. Low VWF levels can increase VEGF activity and signaling [7].

Angiodysplastic vessels generate more shear stress than normal vessels, and the absence of functional VWF in aortic stenosis patients increases their susceptibility to GIB. However, the presence of angiodysplasia does not always lead to bleeding; some patients who experienced GIB from angiodysplasia prior to aortic valve replacement were found to have angiodysplasia during follow-up endoscopic examinations still, yet their bleeding had ceased. This indicated that bleeding may be more closely associated with the shear stress-induced loss of high-molecular-weight VWF in certain patients. The acquired reduction in high-molecular-weight VWF multimers increases the risk of bleeding, especially in the GI tract [7].

4.3. Risk Stratification and Prevention of Gastrointestinal Bleeding in Valvular Disease

Due to the significant risk of GIB in patients living with valvular diseases, identification of high-risk patients must occur before interdisciplinary mitigation strategies can be implemented to prevent GIB. Desai et al. found that patients with Heyde syndrome have a higher inpatient mortality rate when hospitalized for GIB, stressing the need for risk stratification [26]. The American College of Chest Physicians recommends the use of scoring tools, including the HAS-BLED, HEMORR2HAGES (hepatic or renal disease, ethanol abuse, malignancy, older, reduced platelet count/function, hypertension, anemia, genetic factors, excessive fall risk, and stroke), ATRIA, and ORBIT, to identify high-risk patients. Among these, the HAS-BLED score is particularly noted for its ease of use and relatively higher sensitivity in predicting major bleeding events in patients with atrial fibrillation. The HAS-BLED score includes factors such as hypertension, abnormal renal/liver function, stroke, bleeding history, labile INR, age over 65, and concomitant use of drugs or alcohol [52]. In addition, the Academic Research Consortium for High Bleeding Risk (ARC-HBR) created a validated set of criteria to identify high-bleeding risk patients, particularly in the context of percutaneous coronary intervention. According to the ARC-HBR, patients are considered to be at a high bleeding risk if they meet at least one major criterion or two minor criteria. Major criteria include severe anemia, thrombocytopenia, and recent major bleeding, while minor criteria include age ≥ 75 years and moderate anemia [53].

Although no explicit scoring tools specific for patients with Heyde syndrome have been validated, special attention should be paid to those with high-risk features. Caspar et al. and Blackshear et al. demonstrated a direct correlation between the severity of aortic stenosis and mitral regurgitation, respectively, to the degree of acquire von Willebrand syndrome, such that that those with severe disease may be at a higher risk of GIB [6,15]. None of the available studies evaluated the relationship between the severity of tricusupid regurgitation and the risk for GIB, which poses an opportunity for future research. The multicenter retrospective study by Sugino et al. suggested that patients with moderate to severe anemia and severe aortic stenosis demonstrated significantly more angiodysplasia throughout the gastrointestinal tract and active bleeding compared to those without [18]. Furthermore, older age, significant uremia, or histories of anemia and/or GIB should increase the suspicion for Heyde syndrome and should be carefully monitored for signs of bleeding [9,19,45]. Evidence has also suggested that dysfunctional valve prostheses or residual paravalvular leak following valvular repair or replacement is associated with recurrence GIB, commonly due to angiodysplasia [43,46].

Several noninvasive lifestyle changes and medical management strategies can be implemented to reduce GIB risk in patients with valvular disease. Avoidance of medications that make patients more prone to bleeding, like non-steroidal anti-inflammatory agents, especially in elderly patients or those on systemic anticoagulation, is crucial [54]. Adjunct proton pump inhibitors can also be beneficial, especially for antithrombotic or anticoagulation therapy patients. This drug class has been shown to reduce the risk of GIB in patients with valvular disease [55,56]. Addressing comorbid conditions like Helicobacter pylori infection and peptic ulcer disease can reduce GIB risk [54]. Interestingly, a study involving patients with left ventricular assist devices suggested that a diet risk in omega-3 fatty acids may also help reduce the risk of GIB [57].

4.4. Management of Gastrointestinal Bleeding in Valvular Disease

In patients with symptoms of GIB (e.g., melena, hematochezia, iron deficiency anemia), aortic stenosis (e.g., dyspnea, angina, syncope), or other valvular disturbances like mitral or tricuspid regurgitation, a diagnosis of Heyde syndrome should be considered. A transthoracic echocardiogram can be performed to confirm the presence and severity of the valvular disease. Laboratory tests can evaluate signs of anemia or coagulopathy. Specific tests can include complete blood counts to assess for anemia, coagulation profiles with a prothrombin time and activated partial thromboplastin time, and VWF assays, including VWF antigen, VWF ristocetin cofactor activity, and multimer analysis to detect loss of high-molecular-weight VWF multimers. In cases where standard VWF assays are normal but clinical suspicion remains high, platelet function tests, such as closure time with collagen/ADP and collagen/epinephrine, can be considered to detect defects in platelet aggregation under high shear stress [58,59,60]. Finally, upper and lower endoscopy can be performed to confirm or identify sources of GIB, particularly angiodysplasia [61].

If Heyde syndrome is confirmed, evidence has shown that valve replacement can improve GIB and normalize VWF multimers [50]. Both surgical and transcatheter aortic valve replacement have been shown to be effective in resolving the bleeding diathesis from Heyde syndrome. A meta-analysis by Goltstein et al. reported a recovery of acquired von Willebrand syndrome in a significant portion of patients with GIB (73% of cases) [27]. This was further supported by another Mayo Clinic study that reported that 79% of patients experienced no further GIB after AVR, particularly with bioprosthetic valves [62]. TAVI has also been effective in cases where it was shown to have an 88% resolution of GIB in patients with Heyde syndrome [63]. Furthermore, Goltstein et al. reported that GIB ceased in 62% of such patients within the first year post-TAVI and showed further improvement over time [16].

The AVR approach depends on the patient’s surgical candidacy and other comorbidities. Several studies have been published on the incidence of GIB after AVR with a transcatheter approach typically performed in higher-risk patients. Brown et al. found that TAVI may be effective and safe for patients with comorbid severe AS and GIB, as 79.3% had resolution of GIB after TAVI [23]. As TAVI procedures are becoming more commonplace, an increasing rate of next-day discharges has been seen at the expense of increasing rates of GIB [24]. Moreover, the retrospective study of Iyengar et al. showed that patients with TAVI have more frequent readmissions for late GIB than patients with SAVR [25]. Conversely, Desai et al. found that TAVI patients may have fewer complications than the SAVR, with no significant difference in all-cause mortality, indicating that TAVI could be a promising alternative for patients with Heyde syndrome [26]. This further reinforces the intricate nature of multidisciplinary discussion on deciding the modality of surgical intervention for severe AS. The type of SAVR can further increase the risk of bleeding, mainly in patients with mechanical valves, adding further complications to the decision-making process before AVR [28].

One study by Cai et al. suggested GIB related to severe tricuspid regurgitation could be effectively treated with adjunctive transcatheter tricuspid valve implantation in addition to goal-direct medical therapy [5]. Blackshear et al. found that even mitral valve replacement may reverse acquired von Willebrand syndrome and lead to lower rates of GIB [6]. As Fukunaga et al. and Rokui et al. demonstrated, GIB rates are similar with mechanical or bioprosthetic valves [40,41]. However, mechanical valves show superior survival and freedom of thromboembolic events and, therefore, should be preferred [40]. Percutaneous valvuloplasty can also be used to definitively treat Heyde syndrome related to the mitral valve [46]. However, Meindl et al. refuted these results, showing that transcatheter mitral valve repair does not improve VWF levels or VWFAct four weeks after the procedure [35]. However, residual paravalvular leaks in both valve replacements and repairs have been associated with recurrent GIB, most commonly due to angiodysplasia, suggesting improper valve placement or inappropriately sized replacement valves may potentiate Heyde syndrome [43]. Grasso et al. suggest GIB may even be a long-term complication of valvuloplasty; however, the pathophysiology is less understood [38]. Conversely, Liu et al. found no difference in the age-related incidence of GIB between patients who underwent valvuloplasty versus valve replacement [50]. Comorbid pulmonary hypertension may also increase the risk of GIB in patients undergoing transcatheter mitral valve repair [39].

The criteria and guidelines for identifying patients needing valve replacement for aortic stenosis, mitral regurgitation, or tricuspid regurgitation are well defined by several professional societies, including the American College of Cardiology (ACC) and the American Heart Association (AHA). For aortic stenosis, the ACC/AHA guidelines recommend aortic valve replacement in patients with severe aortic stenosis who are symptomatic, regardless of surgical risk, or in asymptomatic patients with left ventricular ejection fraction (LVEF) < 50%. Severe aortic stenosis is defined by an aortic valve area ≤ 1.0 cm², a peak aortic jet velocity ≥ 4.0 m/s, or a mean transvalvular gradient ≥ 40 mmHg. For mitral regurgitation, the ACC/AHA guidelines recommend surgery in symptomatic patients with severe mitral regurgitation and LVEF > 30% or asymptomatic patients with severe mitral regurgitation and LVEF 30–60% or left ventricular end-systolic dimension ≥ 40 mm. For primary tricuspid regurgitation, any signs and symptoms of right-sided heart failure and severe primary tricuspid regurgitation may benefit from isolated tricuspid valve surgery to reduce symptoms and recurrent hospitalizations. Tricuspid valve surgery can be beneficial if tricuspid annular dilation (end-diastolic diameter > 4.0 cm) or prior signs and symptoms of right-sided heart failure are present [64].

The ACC and AHA provide detailed guidelines for anticoagulation in patients undergoing valve surgery. For patients with mechanical valve prostheses, the recommended target INR varies based on the valve position and patient risk factors. For patients with a current-generation mechanical aortic valve and no additional thromboembolic risk factors, the target INR is 2.5 (range 2.0–3.0). This range balances the risks of thromboembolism and bleeding. For patients at higher risk of thromboembolic complications (e.g., atrial fibrillation, previous thromboembolism, hypercoagulable state, older-generation prosthesis), the target INR should be 3.0 (range 2.5–3.5). The incidence of thromboembolism is higher with mitral mechanical valves than with aortic valves. Therefore, a higher INR target is recommended. The target INR for mitral mechanical valves is 3.0 (range 2.5–3.5). For patients with bioprosthetic valves, lifelong anticoagulation is generally not required unless there is an independent indication, such as atrial fibrillation. However, anticoagulation is recommended early after valve implantation to reduce the risk of thromboembolism until the prosthetic valve is fully endothelialized. For bioprosthetic valves, warfarin is typically targeted to an INR of 2.5 (range 2.0–3.0) for the first 3 months post-surgery, often in combination with low-dose aspirin [64].

Despite warfarin generally being preferred and recommended by the guidelines in patients with severe native valve disease or bioprosthetic valves with comorbid conditions requiring systemic anticoagulation, evidence has suggested NOACs may be non-inferior and offer an improved bleeding risk. Yang et al. demonstrated that NOACs have comparable efficacy for ischemic stroke and systemic embolic events and similar rates of major bleeding, including GIB, in patients with tricuspid regurgitation [43]. Briasoulis et al. found that dabigatran and rivaroxaban were associated with significantly lower risk of death in patients with valvular heart disease with atrial fibrillation and were comparable to warfarin for ischemic stroke and GIB rates [46]. Duan et al. and Suppah et al. found NOACs to be non-inferior to warfarin in thromboembolic prevention in patients with atrial fibrillation and bioprosthetic valves but should be preferred given their more favorable adverse bleeding effect profile [48,49].

Endoscopic management of GIB from angiodysplasia related to Heyde syndrome is often used for immediate hemostasis. Techniques such as argon plasma coagulation can be effective for accessible lesions, although the recurrence rate of bleeding remains high unless definitive therapy with valve replacement can be performed [13,59,65]. The American Gastroenterological Association guidelines also support the use of valve replacement in Heyde syndrome, highlighting its role in reducing bleeding risk [61]. Pharmacological treatments include using somatostatin analogs like octreotide, which have shown efficacy in reducing bleeding episodes in patients with gastrointestinal angiodysplasia. Additionally, anti-angiogenic therapy with systemic bevacizumab has been explored as a salvage therapy for persistent bleeding post-AVR, with promising results in case reports [66,67,68].

5. Conclusions and Future Directions

Gastrointestinal angiodysplasia is a significant vascular anomaly characterized by dilated, tortuous blood vessels in the gastrointestinal tract involving chronic, intermittent obstructions of submucosal veins, leading to increased VEGF-dependent proliferation and subsequent vessel dilation. This systematic review not only demonstrates the association between angiodysplasia and aortic stenosis but also highlights mitral regurgitation and tricuspid regurgitation as potential etiologies. Definitive management with valvuloplasty or valve replacement is vital to preventing the onset or recurrence GIB in patients with valvular disease. Future research should delve deeper into the relationship between tricuspid and mitral valvular diseases and the development of angiodysplasia and, consequently, GIB. Studies should investigate the effect of tricuspid regurgitation on acquired von Willebrand syndrome and its relationship with GIB. Addressing valve-specific angiodysplasia with targeted management strategies could significantly enhance patient care and outcomes in those affected by this complex vascular condition.