In the United States, cardiovascular disease (CVD), specifically acute myocardial infarction (AMI) and stroke, is the leading cause of disability and mortality [1]. Aspirin is widely accepted as a secondary CVD prevention after AMI or stroke [2]. Moreover, in large surveys, use rates for primary CVD prevention in the population increased steadily in both men and women from 1980 to 2009, reaching approximately 21% of men and 12% of women aged 25 to 74 years [3].

However, the conflict between the benefits and harm of aspirin usage persists. Low doses of aspirin can inflict injury to the gastrointestinal (GI) mucosa, with the potential for considerable morbidity and mortality [4]. The primary adverse effect of aspirin is major bleeding, typically defined as bleeding from the GI tract or other sites that require hospitalization or transfusion. In most patients who are on aspirin, bleeding most commonly occurs in the GI tract [5].

Proton pump inhibitors (PPIs) are useful for the prevention of ulcers secondary to low-dose aspirin and nonsteroidal anti-inflammatory drug (NSAID) use [6–8]. Additionally, a previous study showed that mucoprotective agents (MPAs) such as misoprostol can decrease the risk of upper GI bleeding in aspirin or NSAID users [9]. Therefore, MPAs are often prescribed in clinical practice to prevent or treat upper GI bleeding in addition to PPI or histamine-2 receptor antagonist (H2RA). However, relevant evidence for MPAs except misoprostol is limited, and no large-scale studies have been conducted. In addition to upper GI bleeding, the treatment and prevention of aspirin-associated lower GI bleeding remain unclear and difficult, because the pathogenic mechanisms are different and poorly understood. The national claims data could be considered as a real-world evidence in a large population. Thus, we aimed to investigate the additional effect of MPA, especially eupatilin, on total GI bleeding prevention including lower GI bleeding in concomitant users of aspirin and acid suppressants such as PPI and H2RA using the nationwide database of national claims data from the Korean National Health Insurance Service (NHIS).

To the best of our knowledge, using the nationwide database of national claims data, this is the first study to investigate the additional effect of eupatilin as an MPA on total GI bleeding prevention including the upper and lower GI tracts of aspirin plus concomitant acid suppressant users such as PPI or H2RA. This study is representative of the general population, as data were collated from the Korean NHIS database, which encompasses the entire nation. Other possible risk factors for GI bleeding, such as concomitant anticoagulant use and other comorbidities, were statistically analyzed. Our study findings reveal that the concurrent use of acid suppressants only (without eupatilin) in aspirin users showed a decreased GI bleeding incidence compared to aspirin use without GI drugs. Moreover, the use of eupatilin in addition to acid suppressants demonstrated a significantly lower incidence of GI bleeding in aspirin users compared to concurrent users of aspirin and acid suppressants or aspirin-only users. Also, a 3-month or longer use of eupatilin was deemed more effective.

A major adverse clinical outcome associated with the use of aspirin is GI bleeding. Current evidence shows that aspirin increases the risk of lower GI bleeding to a similar extent as that in the upper GI [10,11]. Lower GI bleeding secondary to aspirin use has become an important medical issue because the number of aspirin users has increased recently, and older age is an important risk factor for lower GI bleeding [12]. Acid suppressants such as PPIs and H2RAs are typically associated with a reduced risk of upper GI bleeding in aspirin users [13]. However, these do not confer protective effects against aspirin-induced lower GI injury. Some studies have shown that PPIs and H2RAs are associated with an increased risk of lower GI bleeding [14–16]. In a previous study with a small sample size, misoprostol has shown efficacy in the treatment of aspirin-induced intestinal injury [17]. However, frequent misoprostol side effects are dose-dependent cramping, abdominal pain, and diarrhea. In many patients, these side effects interfere with compliance [18]. Moreover, effective medications to prevent aspirin-induced lower GI bleeding is not yet available.

PPIs and H2RAs are well-known, effective agents for treating or preventing upper GI bleeding in aspirin users [19,20]. Despite the wide use of PPIs, long-term PPI use has been associated with several safety concerns. Previous studies have reported that PPI use has been associated with an increased risk of mineral malabsorption, bone fracture, and enteric infections [21–24]. The pathophysiologic mechanism involved in the increased risk of infection remains unclear, although the mechanism of action of PPIs is to block the gastric acid secretion in the stomach. Because of the concerns regarding the long-term safety of PPIs, clinicians consider alternative medications such as MPAs. MPAs, including eupatilin, rebamipide, and misoprostol, do not inhibit gastric acid secretion. Eupatilin is the standardized extract of Artemisia asiatica and has been widely prescribed for treating or preventing gastric mucosal lesions in Asia. Eupatilin has been found to have anti-inflammatory and protective effects against GI lesions in both animal models and humans, which were small sample-sized studies [25–28]. A recent large-scale retrospective observational study has found that eupatilin did not reduce the risk of occult GI bleeding in NSAID users [29]. The study used a ≥ 2 g/dL drop in hemoglobin level to define occult GI bleeding and included participants who underwent routine asymptomatic laboratory follow-ups, which were not conducted on patients with overt GI bleeding and could be a study limitation. Conversely, the present study evaluated the overt GI bleeding risk in aspirin users with the concomitant use of MPA, especially eupatilin. Our results revealed that the use of eupatilin in addition to acid suppressants showed a reduced incidence of GI bleeding in aspirin users compared to concurrent users of aspirin and acid suppressants or aspirin-only users.

A previous small bowel capsule endoscopy (CE) registry-based Korean study, including data from 4,650 CE examinations, revealed that patients in whom the indication for CE was overt, obscure GI bleeding accounted for 53.1% of those who had taken NSAIDs, and the rate of CE diagnosis of NSAID enteropathy was 5.9%. Although the study was limited by a small cohort, the number of diagnoses of NSAID enteropathy had increased, and this increase could be attributed to the aging population [30]. Our study group, using the nationwide database of claims data from the NHIS, found that MPAs had a protective effect against small bowel bleeding in patients using aspirin [31]. In that study, patients with anemia, melena, or hematochezia that occurred within 4 weeks before and after CE were suspected to have small bowel bleeding. Overall, 15,542 patients taking aspirin were included in the study, and small bowel bleeding was confirmed in 126 (0.8%). In that study, PPI use (HR, 2.85) was significantly associated with small bowel bleeding, whereas eupatilin use (HR, 0.35) appeared to be preventive. Small bowel bleeding occurred more frequently in patients who used acid suppressants concurrently with PPIs than in those who did not use acid suppressants concurrently with PPIs (1.3 vs. 0.5%). Subgroup analysis revealed that eupatilin significantly reduced the risk of small bowel bleeding in patients using aspirin concurrently with acid suppressants (HR, 0.23 vs. 2.55) [31]. However, the numbers of patients taking aspirin and those with small bowel bleeding were relatively small. Because the CE findings could not be identified from the claims data, a proportion of CE examinations may not have revealed any bleeding origins in the small bowel. To overcome the study limitations, such as a small number of participants and only patients with small bowel bleeding found on CE examinations, we tried to obtain information from adequate number of patients with total GI bleeding using large-scale nationwide population data.

Because CE is now available to detect small intestinal lesions, small intestinal mucosal injuries have been recognized as a common complication associated with NSAIDs. Small intestinal injury does not occur via an acid-dependent mechanism but because of various factors, such as enteric bacteria; bile acids; prostaglandin deficiency; and topical factors, such as abnormal intestinal mucosal permeability, mitochondrial dysfunction, reactive oxygen species, and stress on the endoplasmic reticulum [32]. Several experimental and clinical studies have revealed the effectiveness of MPAs, such as eupatilin in preventing GI lesions [28,33,34]. In a recent meta-analysis based on 12 studies with 341,063 participants, PPI use was associated with an increased risk of lower GI bleeding, particularly in the small bowel. This association was particularly pronounced among patients taking aspirin or NSAIDs [35]. In a previous study, the following strategies against NSAID-induced GI injuries were recommended [32]. In patients at high risk of upper GI disease, such as peptic ulcer, a PPI for upper GI disease and an MPA for the small intestine should be simultaneously administered to prevent NSAID-induced GI injury. In other patients, an effective MPA is adequate to protect the entire digestive tract, from the esophagus to the small intestine [32]. Our results are the first to demonstrate the positive effects of eupatilin as an MPA on the prevention of total GI bleeding based on large population data in a real-world setting, outside the strictly controlled limits of a clinical trial.

This study has limitations. First, because the esophagogastroduodenoscopy, colonoscopy, and abdominal computed tomography findings could not be identified from the claims data, we considered that the bleeding symptoms such as melena and hematochezia were sufficient measures for overt GI bleeding. The causes and locations of GI bleeding were not identified in detail. Nevertheless, our results included an analysis of the risk for total GI bleeding; however, differentiating between upper and lower GI bleeding was not conducted, which is a limitation. Second, we could not distinguish between the patients who took PPIs and those who took H2RAs, although these drugs have different levels of potency and different effects on acid suppression. These acid suppressants can be prescribed in different amounts, such as PPIs in half or full doses and H2RAs once or twice a day; however, such regimens could not be identified and analyzed in detail from the claims data. Despite these limitations, we believe that the present study has value because the protective effect of eupatilin as an MPA on total GI bleeding in aspirin plus acid suppressant users were evaluated based on the nationwide database of claims data.

In conclusion, eupatilin showed an additional preventive effect for GI bleeding occurrence in aspirin users with concomitant use of an acid suppressant. Moreover, overt GI bleeding reduction was more effective with a 3-month or longer eupatilin use. While acid suppressants are widely used for the prevention of GI bleeding in aspirin uses, the coadministration of eupatilin could be expected to decrease GI bleeding risk. A 3-month or longer-term cotreatment of eupatilin could be recommended for reducing GI bleeding in aspirin plus acid suppressant users.