Analysis of the Clinical Characteristics and Endoscopic Features of Phytobezoar-Induced Ulcers and Gastric Ulcers: A Single-Center Retrospective Study in China

Introduction

The term “bezoar” refers to concretions formed from undigested or partially digested material that have built up within the gastrointestinal tract [1]. Although the stomach is the primary site for their formation, bezoars can also develop in other regions, including the esophagus, duodenum, small intestine, and colon [2]. The reported incidence of bezoars is 0.4–1.0% [3]. Based on their constituent materials, bezoars are classified into several types. These include phytobezoars, which are primarily formed from indigestible plant fibers; trichobezoars (Rapunzel syndrome), composed of hair; lactobezoars, made up of milk components like lactose and casein; pharmacobezoars, resulting from accumulated medications such as aspirin or sucralfate; and bezoars formed from various foreign bodies [2,4]. Phytobezoars are the most common type of bezoars [4].

Previous studies have indicated that the formation of bezoars is closely associated with gastric functional abnormalities. Impaired gastrointestinal motility and delayed gastric emptying, as well as hypochlorhydria or achlorhydria, may increase the risk of bezoar formation and subsequently predispose patients to gastric ulcer development. Bezoars may cause abdominal pain, nausea, and vomiting, and in severe cases, lead to gastric mucosal injury and ulceration, potentially exacerbating gastroesophageal reflux symptoms. However, gastroesophageal reflux disease does not necessarily occur in patients with bezoars, and the causal relationship between these conditions requires further clarification and literature support [5,6]. The severity of bezoar-induced ulcers largely depends on the bezoar’s location, type, size, and duration of disease. Clinically, bezoars may cause pyloric obstruction or intestinal obstruction and result in pressure-related mucosal ulceration. As the condition progresses, severe complications such as gastrointestinal perforation and bleeding may occur [6,7]. Therefore, our study aims to compare these 2 types of ulcers in terms of demographic and clinical characteristics, endoscopic features, complications, and treatments, with the aim of reducing misdiagnoses or missed diagnoses of bezoars, preventing complications, and providing optimal treatment plans.

Typically, the treatment of gastric ulcers involves a combination of acid suppressants and protective agents for the gastric mucosal system, and the course of treatment is 6–8 weeks. However, there is no unified standard for the treatment of bezoar-induced ulcers. Our study will aid in standardizing the treatment of bezoar-induced ulcers based on the similarities and differences between the 2 diseases.

Material and Methods

PATIENTS:

This study is a single-center retrospective analysis. From October 2014 to April 2025, 247 of the 59 773 patients (0.413%) who underwent endoscopy (using a GIF-HQ290 or GIF-Q260 endoscope, Olympus Optical Co., Ltd., Tokyo, Japan) in our medical center were diagnosed with bezoars. In addition, 26 patients were diagnosed with bezoars by abdominal computed tomography (CT) without endoscopy. All together, 163 patients with bezoar-induced ulcers among 273 cases of bezoars were selected as Group A, and 155 patients with gastric ulcers confirmed by endoscopy were selected as Group B, concurrently. Patients with malignant ulcers, confirmed by pathology, were excluded from the study.

For the patients in Group A, most of the patients with phytobezoars were treated with endoscopic lithotripsy, which employed a Shanghai WILSON Bezoars Cutting Loop Sleeve to cut the bezoars into pieces, combined with drug lithotripsy. Drug lithotripsy was performed by giving the patient 5% sodium bicarbonate (250 mL or 500 mL/day) orally or via nasogastric lavage injection along with intravenous administration of proton pump inhibitors (PPIs) and glucose in the initial stage of the study (about 5 years). On the basis of the early data, we retrospectively found that the efficacy of drug lithotripsy was comparable to that of endoscopic lithotripsy combined with drug lithotripsy. However, drug lithotripsy as sole treatment had a lower average hospitalization period, average hospitalization cost, second endoscopy rate, and average endoscopic operation time, along with higher patient tolerance [5]. Therefore, we chose drug lithotripsy alone in the later stage of the study (about 6 years), with endoscopic lithotripsy added only when drug lithotripsy failed after 2–4 days.

Moreover, patients were asked to fast for 2–6 days. They were then allowed to eat after the phytobezoars were dissolved but told to take acid suppressants, PPIs, and the gastric mucosal protective agent rebamipide, once the absence of bezoars and intestinal obstruction was confirmed by gastroscopy and abdominal CT. The patients in Group B were typically treated with PPIs and gastric mucosal protective agents alone.

STUDY DESIGN:

This study received approval from our institution’s internal review board (IRB-AF-37-02), and informed consent was obtained from all participating patients. We collected and analyzed data related to demographic characteristics, previous history (medical and surgical history, disease), symptoms, size and location of bezoar-induced ulcers and gastric ulcers, complications, and therapy methods.

STATISTICAL ANALYSIS:

Data analysis was performed with IBM Statistical Package for Social Sciences (SPSS), Version 26. Continuous variables, expressed as median (range) or mean±standard deviation, were compared using independent t-tests. Categorical variables, presented as frequencies and proportions, were analyzed using chi-square (χ²) tests. A P-value of less than 0.05 was defined as the threshold for statistical significance.

Results

BASELINE CHARACTERISTICS AND CAUSATIVE FACTORS:

In Group A, the patients had a median age of 63 years (range: 24–90 years). This group consisted of 46 male and 117 female participants, yielding a male-to-female ratio of approximately 1: 2.54. Conversely, in Group B, the median age was also 63 years (range: 28–88 years), with 113 male and 42 female patients, resulting in a male-to-female ratio of about 1: 0.37. Statistical analysis revealed no significant difference in age between the 2 groups (t=−.542, P=0.588), while a statistically significant difference was found in sex distribution (P<0.001). In terms of onset time, the incidence rate in Group A during autumn and winter was significantly higher than that in Group B in the current study (autumn and winter: 82.2% vs 56.1%) (P<0.001). The details are shown in Table 1.

Univariate analysis results showed that the common causative factors that were significantly different between Group A and Group B included diabetes mellitus status (41.1% vs 25.2%, P=0.003), smoking history (16.6% vs 47.1%, P<0.001), drinking history (13.5% vs 36.8%, P<0.001), and nonsteroidal anti-inflammatory drug (NSAID) use (20.9% vs 31.6%, P=0.029). The details are shown in Table 1. However, binary logistic regression analysis results showed that only patient sex and seasonal distribution were significantly different between the 2 groups (P<0.005). Binary logistic regression analysis indicated no significant difference in diabetes, smoking consumption, alcohol consumption, and NSAID use between the 2 groups. The details are shown in Table 2.

CLINICAL CHARACTERISTICS AND COMPLICATIONS:

There were some asymptomatic patients in both groups (1.23% vs 14.8%). These patients were most often diagnosed during endoscopy. The common symptoms in the 2 groups included abdominal pain (92.0% vs 35.5%), abdominal distension (47.2% vs 25.1%), nausea (68.1% vs 23.9%), vomiting (43.6% vs 15.5%), acid reflux (55.8% vs 27.8%), and heartburn (45.4% vs 23.3%). These symptoms were more common in Group A. Melena (2.5% vs 36.1%) was more common in Group B (P<0.001 for all). The most common complication in Group A was gastric outlet obstruction (87/163, 53.4%), followed by gastric mucosal erosion (20/163, 12.3%) and intestinal obstruction (11/163, 6.7%). The most common complication in Group B was upper gastrointestinal bleeding (61/155, 39.4%). The details are shown in Table 3.

ENDOSCOPIC DATA:

Among all the patients with bezoars, 93 patients (93/273, 34.1%) were found to have round or ovoid masses with air bubbles and a mottled appearance ranging from 1.5 to 12 cm in diameter via endoscopy, as shown in Figure 1A, 1B. The most common location of bezoars was the gastric antrum (127/273, 46.5%), followed by the gastric body (98/273, 35.9%). A total of 163 patients (163/273, 59.7%) presented with bezoar-induced ulcers, which were mostly multiple ulcers (105/163, 64.4%) and more frequently in the gastric angle (110/163, 67.5%). The mean length of the ulcers was 15.86±10.94 mm (2–50 mm), with 50 ulcers (30.7%) having a length ≥20 mm. Their endoscopic morphological features were as follows: the ulcers were mostly relatively regular and round or oval; some larger ulcers showed irregular shapes; and they were relatively shallow, with flat and clear boundaries covered with white coating at the base, as shown in Figures 1C, 1D and 2A.

A total of 155 patients had gastric ulcers. The majority of these ulcers were single ulcers (104/155, 67.1%) and were located in the gastric antrum (102/155, 65.8%). The mean length of the gastric ulcers was 10.44±7.78 mm (2–40 mm), with 19 ulcers (12.3%) having a length ≥20 mm. Their endoscopic morphological features were as follows: they were mostly relatively regular and round or oval; and they were relatively deep and had a base covered with a yellowish-white coating or blood scab, with the surrounding mucosa being inflamed, swollen, or ulcerated, as shown in Figure 1E, 1F.

There were significant differences in the number of ulcers (P<0.001), the mean ulcer length (t=4.892, P<0.001), and the ulcer location (P<0.001) between the 2 groups. The details are shown in Table 4.

TREATMENT AND EFFICACY:

In this study, all 273 bezoars were phytobezoars. A total of 173 cases (63.4%) were successfully eliminated via drug lithotripsy (5% sodium bicarbonate and PPIs), 96 cases (35.2%) were treated with endoscopic lithotripsy and drug lithotripsy, and 4 cases (1.4%) were treated with surgery. Three patients with small bowel obstruction underwent small bowel dissection for foreign body removal, and 1 patient with intestinal obstruction and gastrointestinal perforation underwent open exploration and gastrostomy.

After bezoars were successfully eliminated, 163 bezoar-induced ulcers were treated with PPIs (omeprazole, pantoprazole, or esomeprazole) and gastric mucosal protective agents (rebamipide, teprenone, or gefarnate) for 6–8 weeks. Bezoar-induced ulcers can be healed with a short treatment, as shown in Figure 2B–2F. After 1 to 2 weeks, 62 patients were reexamined by endoscopy. Of these, 9 patients were cured, and 51 patients’ ulcers were obviously reduced.

A total of 155 gastric ulcers were treated with PPIs and gastric mucosal protective agents for 6 to 8 weeks. After 8 weeks, 64 patients were reexamined via endoscopy; 32 patients were cured, and 32 patients presented with obviously reduced ulcers.

Discussion

LIMITATIONS:

This research has certain limitations that should be acknowledged. Primarily, as a single-center retrospective analysis, it is subject to potential selection biases such as unbalanced bias, nonsimultaneous control bias, and clinical data omission bias. Therefore, the findings require validation through future multicenter prospective studies. In addition, other factors exist that could have caused biased data, such as a possible low diagnosis rate of gastric mucosal erosion.

Conclusions

Bezoar-induced ulcers represent the most frequently observed complication of bezoars. In our study, such ulcers showed a higher prevalence among female patients and in the autumn and winter. Additionally, the symptom of abdominal pain was more common in patients with bezoar-related ulcers compared with those with gastric ulcers from other causes. The 2 types of ulcers can be distinguished based on different endoscopic features, with bezoar-induced ulcers being mostly multiple and superficial and able to be healed with a short treatment.