Endoscopic treatment of small bowel hemangiomas via double-balloon enteroscopy: A single-center comparative study

Abstract

BACKGROUND

Previous studies have not clarified a universally acknowledged treatment approach that can eradicate small bowel hemangiomas while minimizing harm to patients. Although studies have indicated that endoscopic interventions have promising treatment effects with minimal invasiveness, no comparison of different types of endoscopic therapies has been conducted.

AIM

To compare the clinical efficacy, safety, and procedural outcomes of lauromacrogol injection, argon plasma coagulation (APC), and endoscopic ligation via double-balloon enteroscopy for the treatment of small bowel hemangiomas, and to identify lesion-specific optimal endoscopic strategies.

METHODS

Seventy-five participants were enrolled in this study. Thirteen patients received conservative treatment, 20 received lauromacrogol injection, 22 underwent APC, 18 underwent endoscopic ligation therapy, and two received hemostatic clip therapy. Key indicators and prognostic information for different treatment methods were compared between conservative-interventional and within interventional therapies. Categorical variables were compared using the χ² or Fisher’s exact test, and continuous variables were compared using the Student’s t-test or Wilcoxon rank sum test.

RESULTS

Compared with conservative therapy, interventional therapy was associated with a lower rebleeding rate. Compared with ligation therapy, lauromacrogol therapy was associated with less intraprocedural bleeding and showed a higher technical success rate and shorter operative duration. Compared with lauromacrogol and ligation therapies, APC was associated with a shorter procedure duration, a higher technical success rate, and less intraprocedural bleeding.

CONCLUSION

Our observational data suggest that lauromacrogol injection may be a more suitable option for raised small bowel hemangiomas, while APC may be preferable for flat lesions.

Key Words: Small bowel hemangioma; Double-balloon enteroscopy; Lauromacrogol injection; Argon plasma coagulation; Endoscopic ligation; Obscure gastrointestinal bleeding

Core Tip: Small bowel hemangioma is a rare but important cause of obscure gastrointestinal bleeding, and the optimal endoscopic treatment strategy remains unclear. This single-center retrospective study is the first to compare lauromacrogol injection, argon plasma coagulation, and endoscopic ligation via double-balloon enteroscopy. Our findings suggest that lauromacrogol injection is safer and more effective for raised lesions, while argon plasma coagulation is the optimal option for flat hemangiomas, providing practical guidance for endoscopists in clinical decision-making.

INTRODUCTION

Small bowel bleeding is relatively rare, comprising only 5%-10% of all patients with gastrointestinal bleeding (GIB)[1]. Small bowel hemangiomas constitute a major type of vascular lesion (type IV)[2], accounting for 40.4% of small bowel bleeding cases[3]. Clinically, it may be solitary or multiple, with the jejunum being the most common site of involvement[4]. Small bowel hemangioma was particularly difficult to detect in the past, and the advent of capsule endoscopy (CE) and double-balloon endoscopy (DBE) has considerably improved the preoperative diagnosis of this disease[5-7]. Nevertheless, previous studies are mostly case series, and have not clarified a universally acknowledged treatment approach that can eradicate the disease while minimizing harm to patients. Although studies have indicated that endoscopic interventions have promising treatment effects with minimal invasiveness, no comparison of the different types of endoscopic therapies has been conducted[8,9].

Endoscopic therapies, including argon plasma coagulation (APC), ligation therapy, and hemostatic clip therapy, have been reported to effectively treat small bowel hemangiomas effectively[10-12]. Additionally, recently emerging polidocanol (lauromacrogol) injections have shown promising clinical value for small bowel hemangioma[13]. Although several endoscopic modalities have been reported to be effective for small bowel hemangiomas, the above evidence is largely derived from case reports. This lack of comparative data limits the ability to determine the relative efficacy and safety of each technique. The present study provides the first head-to-head comparison of these three endoscopic therapies in a relatively large cohort, offering valuable insights into lesion-specific treatment selection.

MATERIALS AND METHODS

Study design and participants

A retrospective study investigating the effect of different treatment methods on the prognosis of small bowel hemangiomas was conducted at the Changhai Hospital. Between August 2008 and March 2019, 3028 obscure patients with GIB with no lesions found on gastroscopy and colonoscopy underwent CE or DBE at the Department of Gastroenterology, Changhai Hospital (Shanghai, China) to identify potential gastrointestinal sources. DBE was performed within 2 weeks if CE failed to provide a clear diagnosis (Figure 1). Participants were excluded from this study if they had other gastrointestinal diseases that may cause GIB or if there were untreated hemangiomas after receiving interventional therapies in our center. Patients who received a combination of interventional therapies were also excluded. Fifty-three (1.8%) patients who were confirmed with small bowel hemangiomas using CE or DBE were finally enrolled in the study. Eighteen patients underwent CE, and 46 underwent DBE.

Figure 1 Flow chart of study patients. GIB: Gastrointestinal bleeding.

Based on the assessment of an attending doctor and patient’s choice, different treatment methods were adopted, and patients were divided into four groups according to treatment methods (conservative therapy, APC, ligation therapy, and lauromacrogol injection) (Figure 2). The following data were collected: Age, sex, body mass index, clinical manifestation (melena, bloody stool, or anemia), history of blood transfusion, history of drug use (antithrombotic drugs and non-steroidal anti-inflammatory drugs), previous history of related diseases (hypertension, diabetes, ulcerative colitis, Crohn’s disease, and hepatitis), history of surgery and trauma, method of DBE insertion, laboratory data (hemoglobin, albumin, and coagulation function), location, number, and size of lesions, treatment process and times, and bleeding recurrence during follow-up. Abnormal coagulation was defined as ≥ 1 of the following: International normalized ratio > 1.2, activated partial thromboplastin time > 36 seconds, prothrombin time > 3 seconds, and platelet count < 10 × 10⁹/L. Blood transfusion refers to the amount of blood transfused within 24 hours before and after surgery. All the patients were followed-up for at least 1 year via telephone or outpatient clinic visits. Symptoms and bleeding recurrence were recorded. Rebleeding was defined as the occurrence of overt hematochezia or melena after initial hemostasis, or a documented decrease in hemoglobin of ≥ 2 g/dL accompanied by a positive fecal occult blood test, with other potential causes of GIB ruled out based on available clinical records.

Figure 2 Different endoscopic treatments methods. A: Endoscopic treatment with lauromacrogol injection; B: Endoscopic treatment with argon plasma coagulation; C: Endoscopic treatment with endoscopic ligation.

CE procedure

All CEs were performed at our center using a MiroCam capsule (MiroCam, Introduction Medic, Seoul, Korea) or an OMOM capsule (Jinshan Science and Technology Company, Chongqing, China). Bowel preparation was performed 12 hours before the examination, including a low-residue diet for 24 hours, ingestion of 4 L polyethylene glycol solution (Jiangxi, Hengkang, Nanchang, China), and an overnight fast. The recording device was removed after 8-hour ingestion.

DBE procedure

DBE was performed using EN-450T5 (before 2013) and EN-580T enteroscopes (Fujinon, Saitama, Japan). For those who had a hemangioma under CE but still required enteroscopy to confirm the diagnosis, proximal or distal DBE was determined according to the lesion site. For patients who did not undergo CE, an anal approach was used first, and oral DBE was scheduled if no obvious lesions were detected via the oral route. Patients undergoing the anal approach must take 2 L of polyethylene glycol solution for bowel preparation and receive pethidine hydrochloride as a sedative. Before the oral DBE procedure, an overnight fast of 12 hours was required, and intravenous anesthesia with tracheal intubation was administered. Experienced endoscopists decided the treatment method according to the facilities at the time.

Lauromacrogol injection

After the hemangioma was identified with DBE via the oral or anal route, endoscopic injection with lauromacrogol (10 mL per ampoule, Shanxi Tianyu Pharmaceutical Co., Ltd., China) was performed using Boston Injection needle (Boston Scientific 23G). The drug was injected into the hemangioma after confirmation by drawing back blood. Injection dose is 1-2 mL for each hemangioma lesion (varies according to lesion size), and multiple injections are possible.

APC therapy

APC was applied to flat lesions using an ERBE ICC 350 (United States Incorporated Surgical Systems, Atlanta, GA, United States) machine. Energy delivery rate was set at 30-60 watts per minute. The tip of the argon coagulation catheter was placed approximately 0.5 cm above the lesion, and non-contact coagulation was performed for 1-3 seconds each time until the lesions were burned off under endoscopy.

Endoscopic ligation therapy

A loop ligation device (PolyLoop; Olympus, Tokyo, Japan) was used around the base of the lesion. After the nylon loop was tightened, the lesion was ischemic, turned purple, and could fall off after 1-2 weeks.

Hemostatic clip therapy

The active bleeding site was identified using enteroscopy. After determining the bleeding site, blood vessels of the bleeding lesion were directly clamped using a hemostatic clip (HX-600-135; Olympus Optical Co. Ltd., Tokyo, Japan).

Conservative treatment

Conservative treatments included iron supplementation, blood transfusions, monitoring of vital signs, water and electrolyte supplementation, blood transfusion, gastric acid and protease inhibition, and fasting. Thalidomide 50-100 mg per day was administered to some patients.

Statistical analysis

All statistical analyses were performed using SPSS 22.0 (IBM, Armonk, NY, United States). Categorical variables were compared using the χ² or Fisher’s exact test, and continuous variables were compared using the Student’s t-test or Wilcoxon rank sum test. All the tests were two-sided, and P < 0.05 was considered statistically significant.

RESULTS

Demographic findings

Seventy-five participants were finally enrolled in this study, including 27 (36.00%) males and 48 (64.00%) females. Baseline clinical characteristics of the patients are shown in Table 1. In our cohort, 14 (18.67%) participants had hypertension, six (8.00%) had hepatitis B viral infection, 16 (21.33%) had abnormal coagulation, and 16 (21.33%) had active bleeding on endoscopy. Average number of lesions was 3.62, and mean age of the 75 participants was 42.95 years (range 14-75 years). More than half (57.33%) of the lesions were located in the ileum, and the jejunum (37.33%) was the second most common area. Thirteen (17.33%) patients received conservative treatment, 20 (26.67%) received a lauromacrogol injection, 22 (29.33%) underwent APC, 18 (24.00%) underwent endoscopic ligation therapy, and two (2.67%) received hemostatic clip therapy. During follow-up, rebleeding occurred in 13 (17.33%) patients.

Table 1 Clinical characteristics of the study patients, mean ± SD/n (%).

Characteristic		Results (n = 75)
Age, years		42.95 ± 19.90
Sex	Male	27 (36.00)
	Female	48 (64.00)
BMI		21.31 ± 3.34
History of blood transfusion		28 (37.3)
NSAID		1 (1.33)
Hypertension		14 (18.67)
Diabetes		3 (4.00)
Family history of hemangioma		0 (0)
Hepatitis		6 (8.0)
History of surgery and trauma		35 (46.67)
History of radiotherapy and chemotherapy		3 (4.00)
Abnormal coagulation		16 (21.33)
Albumin, g/L		39.09 ± 3.69
Hemoglobin, g/L		86.03 ± 21.15
Interventional therapy		62 (82.67)
Method of DBE insertion	Oral	43 (63.24)
	Anal	25 (36.76)
Length of operation, minutes		80.53 ± 27.83
Blood transfusion volume, mL		225.00 ± 440.82
Active bleeding		16 (21.33)
Lesion size, cm		0.99 ± 0.61
Number of lesions		3.62 ± 4.11
Location of lesions	Duodenum	4 (5.33)
	Upper jejunum	18 (24.00)
	Lower jejunum	10 (13.33)
	Upper ileum	22 (29.33)
	Lower ileum	21 (28.00)
Treatment methods	Conservative treatment	13 (17.33)
	Lauromacrogol injection	20 (26.67)
	Argon plasma coagulation therapy	22 (29.33)
	Ligation therapy	18 (24.00)
	Clip therapy	2 (2.67)
Bleeding recurrence		13 (17.33)

BMI: Body mass index; NSAID: Nonsteroidal anti-inflammatory drug; DBE: Double-balloon endoscopy.

Comparison of conservative and interventional therapies

Data comparison between conservative and interventional therapies are shown in Table 2. The control (conservative therapy) and interventional therapy groups were included in the analysis. The interventional therapy group had obvious advantages in preventing rebleeding (9.68% vs 53.85%, P = 0.001), and there was no significant difference between the two groups in important baseline characteristics, such as the number of lesions, lesion size, and percentage of coagulation dysfunction. However, patients in the control group were older (55.38 years vs 40.34 years, P = 0.012), and the proportion of patients with hypertension in the control group notably increased (46.15% vs 12.90%, P = 0.012). In a subgroup analysis, the APC group showed the best therapeutic effect. GIB recurrence rate was significantly lower in the APC than that in the control group (4.55% vs 53.85%). Although patients in the lauromacrogol injection subgroup had larger lesions (1.44 cm vs 0.77 cm, P = 0.031), more lesions (5.05 vs 2.11, P = 0.024), and a higher proportion of surgical history (70.00% vs 30.77%, P = 0.038), the lauromacrogol group showed a significantly lower bleeding recurrence (10.00% vs 53.85%, P = 0.013).

Table 2 Comparison of conservative therapy and interventional therapy, mean ± SD/n (%).

		Conservative therapy (n = 13)	Interventional therapy (n = 62)	P value1	Ligation therapy (n = 18)	P value2	Lauromacrogol injection (n = 20)	P value3	Argon plasma coagulation therapy (n = 22)	P value4
Age, years		55.38 ± 12.29	40.34 ± 20.27	0.012	32.83 ± 8.73	0.001	25.00 ± 8.80	0.001	59.55 ± 19.30	0.491
Sex	Male	7 (53.85)	20 (32.26)	0.204	2 (11.11)	0.017	9 (55.00)	0.728	9 (40.91)	0.503
	Female	6 (46.15)	42 (67.74)		16 (88.89)		11 (55.00)		13 (59.09)
BMI		24.25 ± 3.56	20.70 ± 2.97	0.000	18.78 ± 1.71	0.001	20.74 ± 2.98	0.005	22.13 ± 3.12	0.075
History of blood transfusion		4 (30.77)	24 (38.71)	0.756	12 (66.67)	0.073	11 (55.00)	0.284	0 (0.00)	0.014
Antithrombotic		0 (0)	0 (0)	NA	0 (0.00)	NA	0 (0.00)	NA	0 (0.00)	NA
NSAID		0 (0.00)	1 (1.61)	1.000	0 (0.00)	1.000	0 (0.00)	1.000	1 (4.55)	1.000
Hypertension		6 (46.15)	8 (12.90)	0.012	2 (11.11)	0.043	0 (0.00)	0.002	6 (27.27)	0.292
Diabetes		1 (7.69)	2 (3.23)	0.440	0 (0.00)	0.419	0 (0.00)	0.394	2 (9.09)	1.000
Family history of hemangioma		0 (0)	0 (0)	NA	0 (0.00)	NA	0 (0.00)	NA	0 (0.00)	NA
Hepatitis		3 (23.08)	3 (4.84)	0.061	0 (0.00)	0.064	0 (0.00)	0.052	3 (13.64)	0.648
History of surgery and trauma		4 (30.77)	31 (50.00)	0.238	14 (77.78)	0.013	14 (70.00)	0.038	2 (9.09)	0.166
History of radiotherapy and chemotherapy		1 (7.69)	2 (3.23)	0.440	0 (0.00)	0.419	0 (0.00)	0.394	2 (9.09)	1.000
Abnormal coagulation		2 (15.38)	22 (35.48)	0.722	3 (16.67)	1.000	7 (35.00)	0.263	3 (13.64)	1.000
Albumin, g/L		38.25 ± 5.14	39.26 ± 3.37	0.396	41.06 ± 2.86	0.065	39.65 ± 2.60	0.313	37.64 ± 3.75	0.692
Hemoglobin, g/L		95.23 ± 28.28	84.10 ± 19.06	0.104	85.22 ± 15.43	0.215	90.25 ± 25.75	0.605	79.77 ± 12.18	0.031
Blood transfusion volume, mL		NA	232.26 ± 459.78	NA	200.00 ± 388.06	NA	310.00 ± 451.78	NA	163.64 ± 500.04	NA
Lesion size, cm		0.77 ± 0.88	1.01 ± 0.58	0.362	1.28 ± 0.35	0.046	1.44 ± 0.40	0.013	0.46 ± 0.37	0.202
Number of lesions		2.11 ± 1.27	3.84 ± 4.34	0.241	4.33 ± 5.09	0.213	5.05 ± 3.56	0.024	2.27 ± 4.10	0.909
Location of lesions	Duodenum	1 (7.69)	3 (4.84)	0.985	0 (0.00)	0.420	3 (15.00)	0.945	0 (0.00)	0.442
	Upper jejunum	3 (23.08)	15 (24.19)		3 (16.67)		5 (25.00)		6 (27.27)
	Lower jejunum	2 (15.38)	8 (12.90)		1 (5.56)		3 (15.00)		4 (18.18)
	Upper ileum	4 (30.77)	18 (29.03)		11 (61.11)		4 (20.00)		3 (13.64)
	Lower ileum	3 (23.08)	18 (29.03)		3 (16.67)		5 (25.00)		9 (40.91)
Bleeding recurrence		7 (53.85)	6 (9.68)	0.001	3 (16.67)	0.052	2 (10.00)	0.013	1 (4.55)	0.002

¹Comparing interventional therapy with conservative therapy.

²Comparing ligation therapy with conservative therapy.

³Comparing lauromacrogol injection with conservative therapy.

⁴Comparing argon plasma coagulation therapy with conservative therapy.

BMI: Body mass index; NSAID: Nonsteroidal anti-inflammatory drug; NA: Not available.

Comparison of different interventional therapies

Data comparisons of different interventional therapies are shown in Table 3. Twenty patients received lauromacrogol injection therapy, and 18 patients received ligation therapy for raised lesions. Although the overall age and proportion of females were higher in the ligation group than that in the lauromacrogol group, there were no significant differences in lesion-related indicators between the two groups, including lesion size (1.44 vs 1.28, P = 0.221), lesion shape (100% vs 100%), number of lesion (5.05 vs 4.33, P = 0.615), and location (P = 0.085). The results showed that, compared with the ligation therapy, the lauromacrogol therapy resulted in less bleeding due to endoscopic operation (15% vs 44.44%, P = 0.046), had a higher technical success rate (95% vs 72.22%, P = 0.055), and shorter operative duration (79.75 minutes vs 97.50 minutes, P = 0.06). No significant difference was observed between the two groups in terms of clinical success (95.00% vs 83.33%, P = 0.653) or adverse events (P = 0.567).

Table 3 Comparison of different interventional therapy, mean ± SD/n (%).

		Lauromacrogol injection (n = 20)	Ligation therapy (n = 18)	P value1	Argon plasma coagulation therapy (n = 22)	P value2
Age		25.00 ± 8.80	32.83 ± 8.73	0.009	59.55 ± 19.30	0.001
Sex	Male	9 (55.00)	2 (11.11)	0.033	9 (40.91)	0.055
	Female	11 (55.00)	16 (88.89)		13 (59.09)
BMI		20.74 ± 2.98	18.78 ± 1.71	0.019	22.13 ± 3.12	0.001
History of blood transfusion		11 (55.00)	12 (66.67)	0.522	0 (0.00)	0.001
Antithrombotic		0 (0.00)	0 (0.00)	NA	0 (0.00)	NA
NSAID		0 (0.00)	0 (0.00)	NA	1 (4.55)	0.415
Hypertension		0 (0.00)	2 (11.11)	0.218	6 (27.27)	0.032
Diabetes		0 (0.00)	0 (0.00)	NA	2 (9.09)	0.167
Family history of hemangioma		0 (0.00)	0 (0.00)	NA	0 (0.00)
Hepatitis		0 (0.00)	0 (0.00)	NA	3 (13.64)	0.065
History of surgery and trauma		14 (70.00)	14 (77.78)	0.719	2 (9.09)	0.001
History of radiotherapy and chemotherapy		0 (0.00)	0 (0.00)	NA	2 (9.09)	0.167
Abnormal coagulation		7 (35.00)	3 (16.67)	0.278	3 (13.64)	0.202
Albumin, g/L		39.65 ± 2.60	41.06 ± 2.86	0.121	37.64 ± 3.75	0.004
Hemoglobin, g/L		90.25 ± 25.75	85.22 ± 15.43	0.476	79.77 ± 12.18	0.199
Blood transfusion volume, mL		310.00 ± 451.78	200.00 ± 388.06		163.64 ± 500.04	0.610
Location of lesions	Duodenum	3 (15.00)	0 (0.00)	0.085	0 (0.00)	0.019
	Upper jejunum	5 (25.00)	3 (16.67)		6 (27.27)
	Lower jejunum	3 (15.00)	1 (5.56)		4 (18.18)
	Upper ileum	4 (20.00)	11 (61.11)		3 (13.64)
	Lower ileum	5 (25.00)	3 (16.67)		9 (40.91)
Lesion shape	Flat	0 (0.00)	0 (0.00)	NA	22 (100)	0.001
	Raised	20 (100)	18 (100)		0 (0.00)
Duration of endoscopic operation		79.75 ± 29.36	97.50 ± 26.69	0.060	66.14 ± 21.21	0.001
Bleeding due to endoscopic operation		3.00 (15.00)	8.00 (44.44)	0.046	2.00 (9.09)	0.018
Lesion size, cm		1.44 ± 0.40	1.28 ± 0.35	0.221	0.46 ± 0.37	< 0.001
Number of lesions		5.05 ± 3.56	4.33 ± 5.09	0.615	2.27 ± 4.10	0.098
Tech success		19.00 (95.00)	13.00 (72.22)	0.055	22.00 (100.00)	0.009
Clinical success		18.00 (90.00)	15.00 (83.33)	0.653	21.00 (95.45)	0.446
Adverse event	Fever	2.00 (9.09)	2.00 (9.09)	0.567	2.00 (9.09)	0.760
	Abdominal pain	1.00 (4.55)	1.00 (4.55)		1.00 (4.55)
	Bloating	1.00 (4.55)	1.00 (4.55)		1.00 (4.55)
	Ulcers	0.00 (0.00)	0.00 (0.00)		0.00 (0.00)

¹Comparing lauromacrogol injection with ligation therapy.

²Comparing lauromacrogol injection, ligation therapy, and argon plasma coagulation therapy.

BMI: Body mass index; NSAID: Nonsteroidal anti-inflammatory drug; NA: Not available.

APC was mainly applied to flat lesions. Compared with the lauromacrogol and ligation therapy, the APC therapy had shorter procedure duration (66.14 minutes vs 79.75 minutes/97.50 minutes, P = 0.001), higher technical success rate (100.00% vs 95.00%/72.22%. P = 0.009), and resulted in less bleeding due to endoscopic surgery (9.09% vs 15%/44.44%, P = 0.018). However, lesion size in the APC group was significantly smaller than that in the other two groups (0.46 vs 1.44/1.28, P < 0.001).

Factors associated with rebleeding

Other factors affecting recurrent bleeding were also explored. Participants were divided into two groups (rebleeding group vs non-rebleeding group) shown in Table 4 according to bleeding occurrence. We found that rebleeding tended to occur in patients with larger (1.41 vs 0.92, P = 0.025) and more (7.91 vs 2.83, P < 0.001) lesions. Additionally, the different treatment methods were closely related to prognosis (P = 0.003).

Table 4 Comparison of recurrent bleeding and non-recurrent bleeding, mean ± SD/n (%).

		Non-recurring bleeding (n = 62)	Recurring bleeding (n = 13)	P value
Age		43.79 ± 19.79	38.92 ± 20.73	0.426
BMI		21.24 ± 3.17	21.64 ± 4.17	0.698
Sex	Male	19 (30.65)	8 (61.54)	0.055
	Female	43 (69.35)	5 (38.46)
History of blood transfusion		23 (37.10)	5 (38.46)	1.000
NSAID		1 (1.61)	0 (0.00)	1.000
Hypertension		11 (17.74)	3 (23.08)	0.699
Diabetes		3 (4.84)	0 (0.00)	1.000
Family history of hemangioma		0 (0.00)	0 (0.00)	NA
Hepatitis		5 (8.06)	1 (7.69)	1.000
History of surgery and trauma		28 (45.16)	7 (53.85)	0.761
History of radiotherapy and chemotherapy		2 (3.23)	1 (7.69)	0.440
Abnormal coagulation		13 (20.97)	3 (23.08)	0.741
Albumin, g/L		39.34 ± 3.40	37.85 ± 4.76	0.181
Hemoglobin, g/L		85.55 ± 20.69	91.23 ± 23.73	0.333
Thalidomide		11 (19.64)	0 (0.00)	0.580
Method of DBE insertion	Oral	37 (62.71)	6 (66.67)	0.819
	Anal	22 (37.29)	3 (33.33)
Active bleeding		15 (26.79)	1 (16.67)	0.590
Length of operation, minutes		79.14 ± 27.69	90.63 ± 28.59	0.277
Treatment methods	Conservative treatment	6 (9.68)	7 (53.85)	0.003
	Lauromacrogol injection	18 (29.03)	2 (15.38)
	Argon plasma coagulation therapy	21 (33.87)	1 (7.69)
	Ligation therapy	15 (24.19)	3 (23.08)
	Clip therapy	2 (3.23)	0 (0.00)
Blood transfusion volume, mL		234.82 ± 452.56	133.33 ± 326.60	0.596
Lesion size, cm		0.92 ± 0.58	1.41 ± 0.68	0.025
Number of lesions		2.83 ± 2.47	7.91 ± 7.63	< 0.001
Location of lesions	Duodenum	2 (3.23)	2 (15.38)	0.208
	Upper jejunum	14 (22.58)	4 (30.77)
	Lower jejunum	8 (12.90)	2 (15.38)
	Upper ileum	21 (33.87)	1 (7.69)
	Lower ileum	17 (27.42)	4 (30.77)

BMI: Body mass index; NSAID: Nonsteroidal anti-inflammatory drug; DBE: Double-balloon endoscopy; NA: Not available.

DISCUSSION

Small bowel bleeding is relatively rare, comprising only approximately 5%-10% of all GIB cases[1]. Because of the anatomical features of the small bowel, detection of bleeding using conventional endoscopy is difficult, and the diagnosis of obscure patients with GIB in this part is filled with dilemma. Gastrointestinal vascular lesions include hemangioma, telangiectasia, angiodysplasia, and phlebectasia, and they account for 5%-10% of all benign neoplasms of the small bowel[14,15]. According to the Yano-Yamamoto classification[2], small bowel hemangiomas are categorized as type IV, which is different from angioectasias, Dieulafoy’s lesions, and arteriovenous malformations. Small-bowel hemangiomas can be life threatening because of their frequent recurrence and persistent bleeding.

Macroscopically, small bowel hemangiomas are typically submucosal, purple-to-red, soft, and pedunculated under endoscopy, as described in previous studies[9,11,13,16], which is consistent with our observations. Histologically, hemangiomas are congenital benign vascular lesions that can be classified as capillary, cavernous, or mixed-type according to the size of the vascular channels[17,18]. Cavernous hemangioma has the highest frequency followed by the mixed type[19]. Capillary hemangioma, which consists of tightly packed submucosal capillaries, is usually solitary and may vary in size, ranging from a nodule of a few millimeters to large lesions (up to 10 cm) extending into the intestinal lumen[20]. However, cavernous hemangiomas originate from larger submucosal arteries and veins. As Pera et al[15] indicated, numerous dilated, irregular blood-filled spaces or sinuses lined by layers of endothelial cells were observed in cavernous hemangioma lesions.

Clinically, small bowel hemangiomas are most commonly found in the jejunum[16]. Concerning the initial common symptoms of small bowel hemangioma, the occurrence frequency of iron-deficiency anemia, pain, and intussusception are 41%, 31%, and 13% respectively[21]. At the same time, the clinical manifestations of the above different histological types have respective characteristics[20]. The main symptom of capillary hemangioma is bleeding, which is often slow and insidious. Its large size may cause intussusception and intestinal obstruction. However, cavernous hemangiomas usually present with severe episodes of hematemesis or melena.

Computed tomography and contrast-enhanced computed tomography are fundamental tools for diagnosing lower GIB[1]. However, gastrointestinal hemangiomas, especially small intestinal hemangiomas, are difficult to diagnose preoperatively. With recent advances in endoscopic techniques, CE and DBE can be recommended for a complete investigation of the small bowel[1,9]. Endoscopic ultrasound, an emerging technology, with DBE has also been reported to be effective for the diagnosis of small bowel diseases and makes it possible to confirm small bowel hemangioma[22,23]. In this study, we have not yet tried to use endoscopic ultrasound as a regular inspection method.

According to a review by Hu et al[9], treatment is indicated for small bowel hemangioma with GIB or abdominal pain. Among the various interventions for small bowel hemangiomas, there is no specific consensus regarding the therapeutic indication. With non-surgical endoscopic approaches, surgical treatment is generally regarded as the last resort[1]. Previous studies have demonstrated the feasibility of individual endoscopic techniques for small bowel hemangiomas; however, most are single-arm case reports or small series with inherent selection bias and no comparator arms. By systematically comparing lauromacrogol injection, APC, and ligation within a single-center cohort, our study addresses this gap and provides preliminary evidence for technique selection based on lesion morphology. First, we compared the clinical effectiveness of conservative treatment, endoscopic lauromacrogol injection, APC, and endoscopic ligation therapy in our endoscopy center. In terms of intraoperative and postoperative recurrent bleeding, lauromacrogol injection showed the lowest incidence. However, compared with the ligation therapy, the lauromacrogol therapy resulted in less bleeding due to endoscopic operation and had a higher technical success rate and shorter operative duration. Based on our single-center experience, ligation often fails because the size of the loop does not match the shape of the hemangioma, which may explain higher operational bleeding rates. We suggested that lauromacrogol injection may be a more safe and effective endoscopic treatment for small bowel hemangioma as previous series case studies reported[13,24-26]. Yet it’s worth noting that although no instances of ectopic embolism or post-injection ulcer bleeding occurred in this cohort, these are recognized potential complications of sclerotherapy. Ectopic embolism may result from inadvertent intravascular injection or excessive sclerosant dosage, while ulceration at the injection site can arise from ischemic injury. Awareness of these risks is essential for procedural safety, and careful injection technique, including aspiration prior to administration and dose titration based on lesion size, should be routinely practiced.

This study has several limitations. First, treatment allocation was non-randomized and based on physician assessment and patient preference, introducing potential selection bias and residual confounding. The baseline imbalances may have influenced both treatment choice and rebleeding risk. Additionally, the lesions in the APC group were smaller, and this baseline difference may have partially contributed to its favorable procedural outcomes. Additionally, due to the retrospective design, repeat endoscopy was not systematically performed in all patients with clinically suspected rebleeding. Therefore, the precise source and mechanism of rebleeding could not be definitively confirmed in a subset of cases. Given these limitations, our findings are better interpreted as indicators of technique suitability for specific lesion profiles rather than as definitive evidence of an optimal strategy.

CONCLUSION

Our results demonstrate the clinical efficacy and tolerability of endoscopic treatments for small bowel hemangiomas. Among these, lauromacrogol injection may be a safer and more effective method. In the future, lauromacrogol injection may have the potential to gradually replace surgery for the treatment of small bowel hemangiomas.