Effect of antithrombotic therapies on small bowel bleeding: an European multicenter retrospective study

Article information

Clin Endosc. 2025;58(1):102-111

Publication date (electronic) : 2024 December 2

doi : https://doi.org/10.5946/ce.2024.073

Lucia Scaramella^,1, Stefania Chetcuti Zammit², Reena Sidhu^3,4, Maurizio Vecchi^1,5, Gian Eugenio Tontini^1,5, Nicoletta Nandi⁵, Matilde Topa⁵, Luca Elli^1,5

¹Gastroenterology and Endoscopy Unit, Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico, Milan, Italy

²Gastroenterology Department, Mater Dei Hospital, Malta, Italy

³Academic Dept of Gastroenterology, Royal Hallamshire Hospital, Sheffield, UK

⁴Department of Infection, Immunity and Cardiovascular Disease, University of Sheffield, Sheffield, UK

⁵Department of Pathophysiology and Transplantation, University of Milan, Milan, Italy

Correspondence: Lucia Scaramella Gastroenterology and Endoscopy Unit, Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico, Via F. Sforza, 35, 20122 Milan, Italy E-mail: lucia.scaramella@policlinico.mi.it

Received 2024 March 26; Revised 2024 June 22; Accepted 2024 July 1.

Abstract

Graphical abstract

INTRODUCTION

Gastrointestinal (GI) bleeding is one of the most important emergencies in gastroenterology and the most common GI diagnosis in the United States necessitating hospitalization, accounting for over half a million admissions annually.1,2

GI bleeding has traditionally been divided into upper GI bleeding (i.e., bleeding in the GI tract between the upper esophageal sphincter and the ampulla of Vater) and lower GI bleeding (i.e., between the ampulla of Vater and the anus), with incidence rates of approximately 47/100,000 and 33/100,000, respectively.3 In the last decade, with the introduction of new techniques to investigate small bowel (SB) pathology, the term SB bleeding (SBB) has been defined as a source of bleeding in the digestive tract between the ampulla of Vater and the ileocecal valve. SBB represents approximately 5% to 10% of all GI bleeding cases, and suspected SBB (SSBB) is the main indication for videocapsule endoscopy (VCE). SSBB can be further classified into overt and occult bleeding, depending on the presence of bleeding manifestations.4-9

Anticoagulants prevent coagulation by inhibiting one or more steps in the coagulation cascade through several mechanisms. This class of drugs includes unfractionated heparin, low-molecular-weight heparin, fondaparinux, vitamin K antagonists (VKAs), and direct factor Xa (FXa) or thrombin inhibitors. Direct oral anticoagulants (DOACs) inhibit thrombin (dabigatran) and coagulation FXa (rivaroxaban, apixaban, and edoxaban). Due to their high safety profile and therapeutic efficacy, they are prescribed more frequently than VKAs in many countries.10,11 Antiplatelet agents (APAs) interfere with the platelet activation process and include several agents such as aspirin (acetylsalicylic acid [ASA], a non-steroidal anti-inflammatory drug [NSAID]), P2Y12 platelet receptor blockers, and platelet glycoprotein (GP) IIb/IIIa antagonists.

The risk of GI bleeding has been a subject of debate since the introduction of DOACs. Recent studies have shown a comparable major GI bleeding risk between DOACs and VKAs in the general population, with a slightly increased risk for DOACs in elderly patients and in the rivaroxaban group,12,13 but accounting for only upper or lower GI bleeding. However, there is a paucity of literature on the impact of DOACs on SBB.

VCE is the diagnostic modality of choice in SSBB because of its high diagnostic yield (DY) (61.7%; 95% confidence interval [CI], 47.3%–76.1%)14,15 and an excellent safety profile, with retention rates of around 1.2% for the indication of bleeding (95% CI, 0.9%–1.6%).16

To date, only a few studies have investigated VCE findings in patients receiving DOACs, suggesting a plausible relationship with SSBB.17-19 In this study, we explored the impact of DOACs on SBB using VCE. Patients receiving anticoagulant/antiplatelet therapies were also included for comparison.

METHODS

Videocapsule endoscopy

Different video capsule prototypes were utilized during the time interval considered and at different sites, such as the mouth to anus (M2A), M2A Plus, PillCam SB2, SB3, Crohn’s Capsule (Given-Medtronic), OMOM (Jinshan Science and Technology), and MiroCam (Intromedic Company). Each investigation, as specified in the European Society of Gastrointestinal Endoscopy (ESGE) guidelines,20 was performed following bowel cleaning with a laxative-based preparation (the schedule varied over time and according to local policy in terms of the type of laxative, volume, and timing of administration) and after an overnight fast. The acquired SB images were analyzed using dedicated software. A prior patency capsule was used if clinically or radiologically indicated.

Indications, use of a patency test, type of capsule, completeness, gastric and SB transit times, SB cleansing, endoscopic findings (type and SB tertile) were collected for each patient and VCE. Cleansing was rated as optimal, good, fair, or poor (dichotomized as adequate when optimal or good, or inadequate when fair or poor). In 2020, the Small Bowel Mucosal Visibility Scoring System was utilized.21

SB findings were classified according to the Saurin classification,22 as suggested by the ESGE guidelines,23 and reviewed at inclusion in the study, considering the recent work by Leenhardt et al.24 (Fig. 1).

Fig. 1.

Saurin classification of small bowel lesions with some examples. (A) Lymphangiectasia. (B) Chylous cysts. (C) Small erosion. (D) Red spot. (E) Angioectasia. (F) Aactive bleeding.

RESULTS

Overall, 1,497 patients (789 [52.7%] males and 708 [47.3%] females; median age, 68 [interquartile range, IQR 58–76] years) underwent VCE. A total of 1756 VCEs were performed. Of these, 1,052 (59.9%) were performed for SSBB (957 patients). The SSBB group was further subdivided into the overt and occult bleeding groups (Table 1).

Table 1.

Demographic characteristics of patients and technical aspects of enteroscopies performed for SSBB (overt and occult) compared with NSBB

The demographic characteristics of the patients and the technical aspects of VCE are reported in Table 1. In the SSBB population, there was no difference in age between females and males (69 [IQR 57–77] years vs. 68 [IQR 60–75] years, respectively; p=0.84). Comparing the overt and occult bleeding subgroups, there was no difference in age (p=0.41) or percentage of patients over 65 years (60.8% vs. 58.6%, p=0.58), whereas there was a slight male predominance in the overt bleeding subgroup (54.7% vs. 46.9%, p=0.046) and a higher completion rate in the occult bleeding group (81.6% vs. 89.0%, p=0.005). Compared with NSBB patients, SSBB patients were older (50 [IQR 37–65] years vs. 68 [IQR 58–76] years, p<0.001), with a male predominance (279 [39.6%] vs. 510 [48.5%], p≤0.001).

The DY was 50.6% in the SSBB group. The most common findings were angioectasia, blood in the lumen, and ulcers (53.9%, 22.3%, and 11.2% respectively; Supplementary Table 1). The DY values of the different subgroups are shown in Supplementary Table 2. Comparing the different DYs, depending on the type of SSBB (overt or occult) or the age of the patient, a greater DY was found in overt SSBB and older patients: 59% (odds ratio [OR], 1.529; 95% confidence interval [CI], 1.126–2.074; p=0.007) and 56.2% (OR, 1.739; 95% CI, 1.357–2.229; p<0.001), respectively (Supplementary Table 3).

A total of 83 patients (8.7%, 37 males) presented with rebleeding episodes (i.e., to a new bleeding episode, regardless of previous treatment) and underwent repeat VCEs. Of the 179 VCEs performed in rebleeding patients, 96 (53.6%) showed positive findings. The most common findings were angioectasia, blood in the lumen, and ulcers (51.8%, 23.4%, and 14.6%, respectively; Supplementary Table 4). In this subgroup, 19 patients (40 VCEs) were on anticoagulant or APAs therapies: one patient (2 VCEs) was taking VKAs, while 18 patients (38 VCEs) were taking APAs. Ten patients (17 VCEs) were treated with ASA alone, one patient was treated with ASA and clopidogrel (2 VCEs) and eight patients (21 VCEs) were treated with clopidogrel. One patient was initially treated with clopidogrel, which was then switched to ASA. Overall, no VCE-related adverse events or capsule retention was reported in this study. Of the patients receiving anticoagulant or antiplatelet therapies, 27 (10.8%) VCEs were performed on those treated with DOACs, 88 (35.2%) on those treated with other anticoagulants (the VKAs subgroup), and 135 (54.0%) on those treated with antiplatelet therapy (the APAs subgroup).

Table 2 shows demographic characteristics of patients and technical aspects of VCE in these patients, showing no differences between the groups in terms of demographic characteristics, completion rate (100%, 89.8%, and 92.6% in the DOACs, VKAs, and APAs groups, respectively) and DYs (48.1%, 48.9%, and 59.3%, respectively). Compared with the general SSBB population, each subgroup consisted mostly of elderly patients, but there were no differences in terms of completion rate and DYs. The most common findings in each group were angioectasia, blood in the lumen, and ulcers (Table 3). Notably, there were no differences in the distinct groups except for diverticula, with a higher frequency in the VKAs group than in the APAs group (4 [6.6%] vs. 1 [0.9%], p=0.0496).

Table 2.

Demographic characteristics of patients and technical aspects of enteroscopies performed for SSBB in patients receiving antithrombotic therapies compared with SSBB population

Table 3.

Videocapsule endoscopic findings in DOAC, VKA, and APA

Subanalysis of this group of patients in two cohorts based on the number of drugs being given (monotherapy vs. combined therapy [i.e. more than one drug]) showed no differences in terms of demographic characteristics and technical aspects of VCE (Table 4) with a DY of 56.0% and 46.9% for monotherapy and combined therapy, respectively (p=0.35). The most common pathologies were angioectasia, blood in the lumen, and ulcers, with no intergroup differences (Table 4).

Table 4.

Demographic characteristics of patients and technical aspects of videocapsule endoscopy and their findings in patients receiving anticoagulants and/or antiplatelet therapies as mono- or combined therapy

We then considered the difference in SSBB patterns according to the variety (Table 5) and number of antithrombotic drugs (Table 6). There was no difference in the overt bleeding rate between anticoagulated (DOACs and VKAs) patients, APAs users, and patients not on antithrombotic therapy, even considering antithrombotic users vs. patients not on therapy (23.2% vs. 20.6%, p=0.59). Similarly, the rate of overt bleeding was similar between the monotherapy and combined therapy groups (17.9% vs. 25.0%, p=0.34).

Table 5.

Suspected small bowel bleeding pattern according to the variety of antithrombotic drugs

Table 6.

Suspected small bowel bleeding pattern according to the number of antithrombotic drugs

DISCUSSION

This study included a large European cohort of patients from two tertiary centers who underwent VCE for SSBB over a 20-year period. It also included a large cohort of patients receiving anticoagulants and/or antiplatelet therapies, in whom SB findings were analyzed and compared to those in the literature.17-19

As previously found in our study,18,26 a high DY in the overt SSBB subgroup was highlighted, presumably because of the intrinsically high DY of VCE in patients with overt bleeding. These results could be ascribed firstly to the ease of localization of overt bleeding lesions or associated with a recent bleeding episode due to the presence of fresh blood or clots that are easily visible, and secondly to the fact that the work up for overt bleeding is usually faster than for occult bleeding with a short time interval between the overt bleeding episode and the execution of the VCE, in accordance with the ESGE guidelines.9

Furthermore, as in our previous study,18 the utility of VCE in the elderly has been established. This could be attributed to the increased number of comorbidities and propensity to develop vascular lesions (e.g., angioectasia), as previously described.27,28

The most common findings, both in the SSBB population and in patients in which rebleeding occurred, were angioectasia, blood in the lumen and ulcers. This finding is consistent with the published literature.18,29,30 The APAs population had a higher number of ulcers than in the rest of the population, but the difference was not statistically significant, and it was the only subgroup with erosions and ulcerated stenosis. Low-dose ASA, was the most common drug in the APAs subgroup, as shown in Table 2. It is largely known that NSAIDs may lead to a drug-induced enteropathy: a clinical entity with a broad spectrum of endoscopic presentations ranging from petechiae and erythematous folds to ulcers and active bleeding.31 Only the most severe findings (e.g., deep ulcers) could be considered as highly relevant bleeding sources, i.e. P2 findings. However, these findings are less frequent, justifying the low DY obtained. In addition, low clinical significance was highlighted in a recent study by Niikura et al.,32 in which the use of NSAIDs, especially low-dose ASA, was significantly associated with an increased risk of SB mucosal injury, but with no significant associations between the use of such drugs and overt SBB.

GI bleeding risk has always been a matter of concern since the introduction of anticoagulants, especially with the advent of DOACs, where comparative studies with VKAs have revealed a higher incidence of GI bleeding associated with DOACs, which is attributable to colon or gastric cancers and diverticular hemorrhage.12,33,34 Evidence in the literature regarding the relationship between DOACs and SBB is sparse and somewhat conflicting. In a descriptive study, Yamaoka et al.19 described a high rate of SB lesions in a small cohort of patients treated with DOACs, whereas Macedo Silva et al.17 concluded that there was a higher risk of SBB in patients receiving DOACs than in patients treated with other anticoagulants and those not on anticoagulants. Another recent study by Choe et al.35 compared acute GI bleeding between DOACs and VKAs and concluded that patients on DOACs had better clinical outcomes in terms of the severity of acute GI bleeding or rebleeding than in those on VKAs. In the SBB subgroup, the frequency was lower in patients taking DOACs (but the difference did not reach statistical significance).

In our previous study, we analyzed the clinical impact of enteroscopy in SSBB, and in a subgroup analysis we concluded that there was no difference in terms of DY between patients receiving DOACs and patients treated with other anticoagulants or antiplatelet therapies.18 In this study, we investigated the possible risk for SBB in patients receiving anticoagulant or antiplatelet therapy. When considering the different drug classes (i.e., DOACs, APAs, and VKAs), the DYs were similar in all groups, and there was no difference when compared with the general SSBB population. The findings were homogeneous within the groups, except for a higher rate of SB diverticula in the VKAs group (4 cases) than in the APAs group (1 case), with no clinical significance.

These results suggest that anticoagulant and antiplatelet drugs are not associated with an increased risk of SBB and do not influence the type of lesion found in the SB. As stated before, the literature is somewhat conflicting; in a recent meta-analysis, Tziatzios et al.36 investigated the impact of antithrombotic treatment on VCE findings in SBB and concluded that antithrombotic treatment was associated with a greater number of positive findings. However, when taking into account the different subgroups, in some studies, anticoagulant drugs were associated with a higher prevalence of P2 lesions, inducing bleeding from SB tumors in particular, and with rebleeding episodes;37,38 however, subsequent larger studies have failed to substantiate this association.39

In relation to APAs agents, the existing literature is also controversial. Boal Carvalho et al.38 found no association between antiplatelet use and VCE positivity, while the meta-analysis of Tziatzios et al.36 concluded that the presence of significant VCE findings was drug-related, and the effect was uniform among anticoagulant and antiplatelet users.

Analysis of the data according to the number of drugs taken by the patients instead of the type showed that the results were in line with prior analyses, with similar DYs and findings, leading to the conclusion that neither the type nor the number of anticoagulant or antiplatelet drugs influenced SB findings or the risk of SBB. The literature is sparse, with only a few studies analyzing the combination of drugs. Two studies40,41 considering only a combination of APAs (low-dose aspirin and thienopyridine) found that this combination significantly exacerbated SB damage, resulting in more erosions and ulcers, while there are no data considering both types of antithrombotic drugs (anticoagulants and APAs). Further larger prospective studies are needed to investigate this relationship.

Rebleeding risk is a matter of concern in patients with SSBB, especially in patients on antithrombotic drugs; some studies have investigated this relationship with contrasting results. However, in their recent work, Tziatzios et al.36 concluded that antithrombotic drugs are a risk factor for rebleeding in these patients.37,42

In our population, only 22.3% of rebleeding cases were in patients on antithrombotic drugs, none were taking DOACs, and only one patient (accounting for two VCEs) was on a VKAs. This is in line with the study by Choe et al.35 which found that rebleeding was less common in patients on DOACs than in those on VKAs. These conflicting results suggest that other confounding factors may influence rebleeding risk; however, further studies are needed.

In our study, we did not find any differences in the SSBB patterns according to the variety or number of antithrombotic drugs used. In literature, there are conflicting results. Boal Carvalho et al.38 found that there was no significant association between the use of either APAs or anticoagulants and an overt SSBB presentation, while Kono et al.43 concluded that patients who took both APAs and anticoagulant drugs had a significantly higher rate of overt presentation in comparison with those on APAs only, but not with those on only anticoagulant drugs. Furthermore, no significant differences were observed between the APAs and anticoagulant groups. In a recent meta-analysis, Tziatzios et al.36 concluded that there was no significant difference between the use and nonuse of antithrombotic therapy for either overt or occult SSBB presentation.

In this study, we confirmed the overall safety profile of VCE, with no documented retention, even in challenging patients, because of the careful use of patency capsules. This completion rate was consistent with the results of our previous study.18

Our study has some limitations, firstly because of its retrospective design and the wide time frames during which patients undergoing VCE were included. This led to the heterogeneity of VCE platforms, with better software and images improving over the years, and reviewers, with different levels of expertise. Owing to the retrospective nature of the study, a large number of patients had missing information, such as an incomplete drug history, which might have affected the results. Additionally, the involvement of only two European centers may limit ethnic diversity, which can influence disease susceptibility and prevalence as well as aspects such as drug response and pharmacogenetics. All these factors can limit the generalizability of the findings to a broader population.

In this study, we confirmed that VCE is a safe technique with a high clinical impact in cases of SBB, especially in challenging situations such as overt SBB, and in the elderly. Furthermore, anticoagulant and antiplatelet therapies do not appear to affect the DY of VCE and, consequently, can be considered safe in terms of the risk of overt SBB.

Anticoagulants and APAs are generally discontinued in patients with acute bleeding. Although more evidence is required from prospective studies with systematic data collection and fewer limiting factors and heterogeneous data (e.g., comprehensive past medical and drug histories to better categorize and stratify the patients, a higher number of patients on antithrombotic drugs, and implementing a standardized follow-up to better detect rebleeding cases), considering the evidence from this study and after weighing the risks and benefits for each patient, it is prudent to restart antiplatelet and anticoagulant use in patients with a history of SBB, as there is no evidence that they increase the SBB risk.

An improved understanding of the impact of antithrombotic drugs in increasing on the risk of SB hemorrhage would lead to better management of these medications in the context of SBB. This would enhance the clinical management of patients and reduce the potential thrombotic risks associated with the inappropriate discontinuation of these drugs, particularly in emergency settings, such as overt SBB.

Supplementary Material

Article information Continued