Open Access
CC BY 4.0 · Journal of Coloproctology 2025; 45(03): s00451810612
DOI: 10.1055/s-0045-1810612
Original Article

Comparative Analysis of Lower Gastrointestinal Bleeding Due to Diverticular Disease Versus Other Etiologies

Authors

  • Camila Vieira Chiquetti

    1   Department of Coloproctology, Regional Hospital of Mato Grosso do Sul, Campo Grande, MS, Brazil
  • Carlos Henrique Marques dos Santos

    2   Brazilian Society of Coloproctology, Rio de Janeiro, RJ, Brazil
  • Bárbara Érnica Zeferino

    1   Department of Coloproctology, Regional Hospital of Mato Grosso do Sul, Campo Grande, MS, Brazil
  • Carlos Otávio da Silva Ribeiro

    2   Brazilian Society of Coloproctology, Rio de Janeiro, RJ, Brazil
  • Odinilson Almeida Fonseca

    2   Brazilian Society of Coloproctology, Rio de Janeiro, RJ, Brazil

Funding The author(s) received no financial support for the research.
 

Abstract

Introduction

Lower gastrointestinal bleeding (LGIB) is defined as bleeding located below the ligament of Treitz, with colonic diverticular disease (CDD) being the main etiology. Characteristics of LGIB caused by CDD were compared with other etiologies to establish the main differences.

Methods

This observational, analytical, and prospective study included individuals with signs suggestive of LGIB, after excluding upper gastrointestinal bleeding via endoscopy. Data were recorded during clinical follow-up from March 2023 to December 2024. A significant level of 5% was considered.

Results

A total of 108 patients was evaluated. The main etiologies were CDD (54.6% - n = 59), hemorrhoidal disease (11.1% - n = 12), angiodysplasia (10.2% - n = 11), and neoplasia (10.2% - n = 11). A significant association was found between CDD and patients over 60 years old (p = 0.035). Patients with CDD sought medical care within two days (p = 0.012), while those in the group with HD, angiodysplasia, and neoplasia sought care after seven days (p = 0.012). Among the patients with CDD, 96.1% were treated clinically, while 65.3% of patients with HD, angiodysplasia, or neoplasia required surgical or endoscopic treatment. HD, angiodysplasia, and neoplasia showed a higher prevalence of hematochezia (58.8%), while CDD showed a higher prevalence of enterorrhagia (71.9%) (p < 0.05).

Conclusion

CDD is the leading cause of LGIB, typically presenting with self-limited bleeding and associated with enterorrhagia. Other causes include hemorrhoidal disease, angiodysplasia, and neoplasms, which often require endoscopic or surgical intervention and present as hematochezia. No significant differences were observed between groups regarding sex, ethnicity, and comorbidities.


Introduction

Lower gastrointestinal bleeding (LGIB) is defined as anorectal bleeding originating below the ligament of Treitz.[1] It is a challenging condition, as the bleeding may stem from a long anatomical pathway, have multiple etiologies, and present clinically with varying characteristics, from severity to stool coloration, often ceasing spontaneously before a definitive diagnosis is made. Some studies have shown that in 10% of all cases, no definitive source of bleeding is identified.[2] [3]

LGIB is a common reason for emergency hospitalization, with an annual rate exceeding 87 cases per 100,000 individuals.[2] [4] Patients and their families often seek urgent medical care due to significant stress and emotional distress caused by the sight of any amount of blood, which may reflect the clinical presentation of severe diverticular bleeding, but may also simply be a symptom of hemorrhoidal disease. Gayer et al.[5] analyzed over 1,100 patients admitted for LGIB and found that more than 20% were diagnosed with hemorrhoidal disease.

The most frequently associated sources of bleeding include colonic diverticula, hemorrhoids, and polyps.[2] Bleeding can occur at any age and manifests as enterorrhagia, hematochezia, or melena,[6] though these associations are not strongly emphasized in the literature.

In most patients, LGIB is a self-limiting condition that resolves spontaneously in 80% of cases, without requiring hospital intervention, and has a mortality rate of 2% to 4%.[2] [7] Despite this, LGIB demands longer hospital stays and more time-consuming and costly examinations until discharge, unlike upper gastrointestinal bleeding (UGIB). A registry of 30,498 hospitalized patients with gastrointestinal bleeding showed that those with LGIB had longer hospitalizations and higher resource use compared to those with upper gastrointestinal bleeding.[8] Clinicians are responsible for identifying which cases involve potentially fatal bleeding that demands hospitalization and the use of critical, costly resources.

Given this context, the objective of the study was to evaluate the main differences between LGIB caused by colonic diverticular disease and other etiologies.


Methods

This study was submitted to and approved by the Research Ethics Committee of the Regional Hospital of Mato Grosso do Sul.

It is an observational, analytical, and prospective study. The selected individuals were all those who presented signs suggestive of LGIB, such as enterorrhagia, hematochezia, and/or melena, and after excluding UGIB through upper gastrointestinal endoscopy, according to institutional protocol. Data was collected from patient history, physical examination, and clinical follow-up. The study was conducted from March 2023 to December 2024, and no intervention was performed.

The following definitions were used to characterize the clinical presentations[9]:

  • Enterorrhagia: elimination of bright red blood through the anus in large quantities, without the presence of feces;

  • Hematochezia: elimination of bright red blood through the anus, generally mixed with feces;

  • Melena: elimination of black, tarry stools with a "coffee ground" appearance and foul odor.

For the evaluation of epidemiological data, the variables sex, age, and ethnicity were considered. The clinical presentation of bleeding, clinical and laboratory characteristics, including all variables present in the Oakland score — were also analyzed. Information was obtained regarding the main bleeding etiologies, defined by endoscopic and imaging exams, and the therapies administered during initial management specific to each cause. Patients were followed throughout hospitalization and after discharge, on an outpatient basis, by the same team.

All collected data were stored in a standardized table using Microsoft Excel® version 2019. Results were analyzed using SPSS software, version 24.0, with a 5% significance level, and presented as descriptive statistics or in tables. The statistical tests employed were the chi-square test (with Bonferroni correction when necessary), the binomial test, and the student's t-test, as described in the table legends.


Results

A total of 108 patients with complaints suggestive of anorectal bleeding were evaluated, of which 57.4% (n = 62) were male (p = 0.149), and the majority were of white ethnicity (53.7% - n = 58). The proportion of patients aged 60 or older (69.4% - n = 75) was significantly higher than that of patients aged 20 to 59 years (p < 0.001). Similarly, the percentage of patients with comorbidities (74.1% - n = 80) was higher than those without (p < 0.001), with the most common comorbidities being systemic arterial hypertension (54.6% - n = 59) and diabetes mellitus (21.3% - n = 23). Most patients had an Oakland score greater than eight (96.3% - n = 104). Of the four patients with a score below eight, none experienced complications such as rebleeding or required clinical intervention or even blood transfusion. Most patients did not have a history of previous LGIB (69.4% - n = 75) but did have blood on rectal examination (74.1% - n = 80, p < 0.001) and did not present with coagulopathy (95.4% - n = 103). Only 31 patients had hemodynamic repercussions—most had isolated tachycardia (17.6% - n = 19), and the rest had hypovolemic shock (11.1% - n = 12). Seventy patients were not on any medications or prior treatments (64.8%). The most used medications were antiplatelet agents (14.8% - n = 16) and anticoagulants (10.2% - n = 11).

Most patients were hospitalized for zero to seven days (75.2% - n = 76), with 18 (16.7%) requiring readmission due to rebleeding. Only five of these were using antiplatelet and/or anticoagulant medications. There was no significant association between the use of these medications and readmission (chi-square test, p = 0.766). Only seven patients (6.5%) died among the 108 treated. For 75.9% of patients (n = 82), treatment was clinical; for 16.7% (n = 18), it was surgical; and for 6.5% (n = 7), endoscopic. The mean hemoglobin value at admission was 8.93 ± 0.23 g/dL, with most patients falling between 7.0 and 10.9 g/dL (66.7% - n = 72). However, most received red blood cell transfusion during hospitalization (59.3% - n = 64). The most frequent clinical presentation was enterorrhagia (59.3% - n = 64), followed by hematochezia (15.7% - n = 17) and melena (10.2% - n = 11). Sixteen patients (14.8%) had more than one type of presentation. The most frequent etiologies were colonic diverticular disease (CDD) (54.6% - n = 59), hemorrhoidal disease (HD) (11.1% - n = 12), angiodysplasia (10.2% - n = 11), and neoplasia (10.2% - n = 11) ([Fig. 1]).

Zoom
Fig. 1 Etiologies of Lower Gastrointestinal Bleeding (LGIB). The results are presented in relative frequency. *Others: colostomy varices; ileofemoral portosystemic shunt; Crohn's disease; fissure; intestinal pseudolipomatosis.

[Tables 1] and [2] show the results of the association between variables and the categorized etiologies into three groups: CDD, HD, angiodysplasia, and neoplasia, and "others." There was a significant association between CDD and age group (p = 0.035), with a higher proportion of patients aged 60 or older having this etiology (61.3% - n = 36), compared to those aged 20–59 (39.4% - n = 13). There was also a significant association between other etiologies and the presence of diabetes mellitus (p = 0.046), with diabetic patients more frequently presenting with other etiologies (30.4% - n = 7) than non-diabetics (12.9% - n = 11). Additionally, there was an association between other etiologies and the use of medications or prior treatments (p = 0.001); patients using anticoagulants or undergoing brachytherapy/radiotherapy were more likely to present with other etiologies (45.5% - n = 5; 100.0% - n = 2, respectively), compared to those not using prior medications or treatments (11.4% - n = 8) (p < 0.05).

Table 1

Results of the evaluation of the association of etiology presented by the patients evaluated in this study with the main variables

Variable

Etiology

HD, angiodysplasia, neoplasm

P-value

DDC

P-value

Other

P-value

Sex

 Male

38,7 (24)

0,060

51,6 (32)

0,465

19,4 (12)

0,384

 Female

21,7 (10)

58,7 (27)

13,0 (6)

Ethnicity

 White

27,6 (16)

0,640

53,4 (31)

0,874

13.8 (8)a

0,030

 Brown

37,2 (16)

58,1 (25)

14.0 (6)a

 Black

20,0 (1)

40,0 (2)

60.0 (3)a

 Yellow

50,0 (1)

50,0 (1)

50.0 (1)a

Age group

 20 to 59

39,4 (13)

0,240

39.4 (13)b

0,035

18,2 (6)

0,779

 60 or +

28,0 (21)

61.3 (46)a

16,0 (12)

Comorbidities

HAS

 No

32,7 (16)

0,811

46,9 (23)

0,143

12,2 (6)

0,261

 Yes

30,5 (18)

61,0 (36)

20,3 (12)

Diabetes

 No

30,6 (26)

0,701

54,1 (46)

0,837

12.9 (11)b

0,046

 Yes

34,8 (8)

56,5 (13)

30.4 (7)a

Cardiopathy

 No

33,0 (30)

0,442

52,7 (48)

0,363

15,4 (14)

0,408

 Yes

23,5 (4)

64,7 (11)

23,5 (4)

Smoking and/or alcoholism

 No

29,7 (27)

0,348

56,0 (51)

0,495

15,4 (14)

0,408

 Yes

41,2 (7)

47,1 (8)

23,5 (4)

Oakland score

 0 to 8

75,0 (3)

0,056

50,0 (2)

0,850

0,0 (0)

0,362

 9 or +

29,8 (31)

54,8 (57)

17,3 (18)

Hemoglobin value at admission (mg/dL)

 Up to 6.9

36,8 (7)

0,425

57,9 (11)

0,950

10,5 (2)

0,230

 7.0 to 8.9

21,6 (8)

51,4 (19)

24,3 (9)

 9.0 to 10.9

34,3 (12)

57,1 (20)

8,6 (3)

 11.0 or +

41,2 (7)

52,9 (9)

23,5 (4)

Medications in use/previous treatments

 No

37,1 (26)

0,123

52,9 (37)

0,155

11.4 (8)c

0,001

 Antiplatelet

37,5 (6)

43,8 (7)

18.8 (3)abc

 Anticoagulant

18,2 (2)

72,7 (8)

45.5 (5)ab

 NSAIDs

0,0 (0)

77,8 (7)

0.0 (0)bc

 Brachytherapy/

 Radiotherapy

0,0 (0)

0,0 (0)

100.0 (2)a

The results are presented in relative frequency (absolute frequency). p-value in the chi-square test.


Different letters in the column indicate a significant difference between the variations of the variables associated with the etiology of the patients in the study (chi-square test with Bonferroni correction, p < 0.05). DDC = diverticular disease of the colon.


Table 2

Results of the evaluation of the association of etiology presented by the patients evaluated in this study with the main variables

Variable

Etiology

HD, angiodysplasia, neoplasm

P-value

DDC

P-value

Other

P-value

Time to seek care (days)

 Up to 2

18.8 (9)b

0,012

68.8 (33)a

0,012

14,6 (7)

0,472

 3 to 7

34.2 (13)ab

52.6 (20)ab

13,2 (5)

 7+

55.5 (11)a

30.0 (6)b

25,0 (5)

Death

 No

28,7 (30)

0,131

56,4 (57)

0,152

15,8 (16)

0,382

 Yes

57,1 (4)

28,6 (2)

28,6 (2)

Length of stay (days)

 0 to 7

27,6 (21)

0,712

52,6 (40)

0,150

14,5 (11)

0,316

 8 to 15

35,0 (7)

75,0 (15)

25,0 (5)

 15+

40,0 (2)

40,0 (2)

0,0 (0)

Type of treatment

 Clinical

15.9 (13)b

<0.001

64.6 (53)a

0,002

14.6 (12)a

0,028

 Surgical

83.3 (15)a

16.7 (3)b

11.1 (2)a

 Endoscopic

85.7 (6)a

42.9 (3)ab

42.9 (3)a

 Other

0.0 (0)ab

0.0 (0)ab

100.0 (1)a

Clinical presentation

 Enterorrhagia

23.4 (15)b

0,020

71.9 (46)a

<0.001

14,1 (9)

0,151

 Hematochezia

58.8 (10)a

29.4 (5)b

29,4 (5)

 Melena

18.2 (2)ab

27.3 (3)b

0,0 (0)

 + that 1

43.8 (7)ab

31.3 (5)b

25,0 (4)

Abbreviation: DDC, diverticular disease of the colon.


The results are presented in relative frequency (absolute frequency). p-value in the chi-square test.


Different letters in the column indicate a significant difference between the variations of the variables associated with the etiology of the patients in the study (chi-square test with Bonferroni correction, p < 0.05).


There was a significant association between the presence of HD, angiodysplasia, or neoplasia and the time taken to seek care (p = 0.012), with 55.5% (n = 11) seeking care after more than seven days, compared to only 18.8% (n = 9) who sought care within two days. Similarly, there was a significant association between CDD and time to seek care (p = 0.012), with 68.8% (n = 33) of CDD patients seeking care within two days, versus 30.0% (n = 6) who waited more than seven days.

To assess the treatments applied, patients with exclusive etiologies for each group were considered. A significant association was found between etiology and the type of treatment received (chi-square test, p < 0.001): 96.1% (n = 49) of CDD patients received clinical treatment, compared to 34.6% (n = 9) of patients with HD, angiodysplasia, or neoplasia. Conversely, 53.8% (n = 14) of patients with HD, angiodysplasia, or neoplasia underwent surgical treatment, compared to 3.9% (n = 2) of CDD patients. Similarly, 11.5% (n = 3) of patients with HD, angiodysplasia, or neoplasia received endoscopic treatment, compared to 0.0% (n = 0) of CDD patients (chi-square test with Bonferroni correction, p < 0.05) ([Fig. 2]).

Zoom
Fig. 2 Chart illustrating the percentage of patients according to etiology and the type of treatment received. Each column represents the percentage value. Each column represents the percentage value. *Significant difference compared to patients with HD, angiodysplasia, and neoplasia for the same type of treatment (chi-square test with Bonferroni correction, p < 0.05).

There was also a significant association between the clinical presentation and the etiology (HD, angiodysplasia, neoplasia: p = 0.020; CDD: p < 0.001). Among patients with HD, angiodysplasia, or neoplasia, 58.8% (n = 10) presented with hematochezia, while 23.4% (n = 15) presented with enterorrhagia. Conversely, 71.9% (n = 46) of CDD patients presented with enterorrhagia, versus 27.3% (n = 3) with melena, 29.4% (n = 5) with hematochezia, and 31.3% (n = 5) with multiple presentations (p < 0.05).

There was a significant association between time to seek care and the clinical presentation (p < 0.001). Among patients who delayed care for more than seven days, 58.8% (n = 10) had hematochezia, significantly more than those with enterorrhagia (6.3% - n = 4) in the same time frame (p < 0.05). In the other time intervals (within two days and between three and seven days), no significant differences were observed between clinical presentations (p > 0.05) ([Table 3]).

Table 3

Results of the evaluation of the association between the time to seek care and the variable clinical presentation

Variable

Time to seek care (days)

P-value

Up to 2

3 to 7

7+

Clinical presentation

 Enterorrhagia

54.7 (35)a

39.1 (25)a

6.3 (4)b

<0.001

 Hematochezia

23.5 (4)a

17.6 (3)a

58.8 (10)a

 Melena

30.0 (3)a

40.0 (4)a

30.0 (3) ab

 + that 1

40.0 (6)a

40.0 (6)a

20.0 (3) ab

The results are presented in relative frequency 9 absolute frequency). p-value in the chi-square test.


Equal letters in the column indicate that there was no significant difference between the clinical presentations (chi-square test with Bonferroni's correction, p > 0.05).



Discussion

Colonic diverticular disease (CDD) is the leading cause of lower gastrointestinal bleeding (LGIB), accounting for at least 50% of such cases.[10] In this population, 54.6% of LGIB etiologies were attributed to this condition. Population-based observational studies, such as that by Hreinsson et al.,[4] reported a lower proportion (23.3%) yet still identified CDD as the main cause of lower intestinal bleeding. Additionally, consistent with the literature, the prevalence of CDD was higher in patients over 60 years of age (p = 0.035). According to Czymek et al[11], McGuire and Haynes[12], and Cheskin, Bohlman, and Schuster[13], diverticular disease has a strong correlation with aging, with approximately 60% of individuals developing diverticula by the age of 80.

In this study, following CDD, the main causes of LGIB were hemorrhoidal disease (HD) (11.1%), angiodysplasia (10.2%), and neoplasia (10.2%), together accounting for 32.5% of cases—a percentage still lower than that of CDD. For statistical analysis, the etiologies were grouped into three categories based on similar incidence: CDD, HD, angiodysplasia, and neoplasia; and "others," which included the remaining etiologies shown in [Figure 1]. It is also important to note that in the neoplasia group, both malignant tumors and polyps were included.

A strong relationship was observed between the etiologies of LGIB, the time taken to seek care, and the clinical presentation of the bleeding. A significant association was identified between hematochezia and neoplasia, HD, and angiodysplasia (p = 0.02), and it was also evident that these patients tended to delay seeking medical attention, with an average delay exceeding seven days (p = 0.012). On the other hand, patients with diverticular disease typically presented with bleeding in the form of enterorrhagia (p < 0.001) and sought medical assistance within two days (p = 0.012), reinforcing the urgent nature of diverticular bleeding.

As expected, many cases (75.9%) were resolved through clinical stabilization alone, without the need for endoscopic or surgical intervention. Specifically, for CDD, this number was even higher in favor of clinical treatment. Indeed, the literature agrees that bleeding from diverticular disease, although often voluminous, is generally self-limited, ceasing spontaneously in up to 80% of cases.[2] [14] [15] In this study, an even higher rate was observed—above 90% ([Fig. 2])—which may be explained by the exclusion, during comparative analysis of treatments, of patients who presented with more than one bleeding etiology in the diagnostic investigation, a methodology not necessarily followed in other studies. It is important to highlight that, at the time of investigation, most patients did not present with active bleeding, which complicated the precise determination of the LGIB origin. The main associated findings were between CDD and ulcers, polyps, or angioectasias. This serves as a warning regarding the risk of presumptive diagnoses of CDD as the cause of LGIB based solely on medical history or imaging (e.g., CT scan), which can lead to delays in more thorough investigation via colonoscopy, even in outpatient settings. Oakland et al.[10] conducted a multicenter audit in the United Kingdom in 2015 and found that, of the 2,528 LGIB cases analyzed, only 29% underwent endoscopic investigation of the gastrointestinal tract. Of these, only half were scheduled for outpatient investigation, raising concerns about the potential for undetected lesions. Furthermore, more than half of the patients readmitted due to rebleeding had not been investigated during their initial admission, suggesting that premature discharge without proper investigation was inadequate.

In this comparative analysis, no significant differences were observed between the main groups in terms of sex, ethnicity, hemoglobin values at admission, or comorbidities. Moreover, no specific studies were found that made this kind of comparison addressing LGIB as a target symptom in the scientific literature. However, sex is well documented in the literature as being associated with colorectal cancer and CDD, with both conditions being more common in men.[1] [5] [16] Likewise, these two diseases are more prevalent in Black individuals compared to white individuals, according to American sources.[16] [17] [18] [19] In contrast, HD appears to affect both sexes similarly, and one study showed its prevalence in white individuals when younger and of higher socioeconomic status.[20]

Regarding the Oakland score, this study had limitations that affected the comparative analysis between LGIB etiologies. The sample of patients with an Oakland score below eight was very small. This may be explained by the regulatory system of the state in which the hospital is located. As a tertiary care hospital, it treats patients referred to from secondary care. In this process, some patients are discharged based on clinical stability, unit protocols, and medical decisions. This patient profile generally has a score below eight, reflecting the natural selection of individuals who do not require prolonged hospitalization or specialized care. Whiteway et al.[21] also found the same limitation, as their study was based on retrospective data collection from the emergency department of East Kent University Hospital Trust, and some patients were treated in another department where data were not recorded, leading to their exclusion from the study.

The use of NSAIDs and antiplatelet agents is well documented in the literature as one of the main risk factors for lower gastrointestinal bleeding, especially in CDD.[2] [22] [23] Hreinsson et al.[4] showed that NSAID use was significantly higher among patients with bleeding (19%) compared to the control group (9%) (P = 0.0096). Likewise, the use of low-dose aspirin was also more common among patients with bleeding (37% vs. 25%, P = 0.0222). Ultimately, this study did not find a significant relationship between prior medications/treatments and bleeding etiologies when compared among themselves. However, 37% (n = 22) of patients with CDD were using antiplatelets/anticoagulants or NSAIDs.

Some data presented in this study could not be compared with the literature due to the scarcity of studies that comparatively address LGIB etiologies, highlighting the need for further research to deepen these findings.


Conclusion

The comparative analysis between the etiologies of LGIB revealed that CDD is the leading cause, generally presenting with self-limited bleeding, while hemorrhoidal disease, angiodysplasia, and neoplasms more frequently require surgical or endoscopic treatment. Differences in the timing of medical consultation and in clinical presentation reinforce the urgent nature of diverticular bleeding, which often manifests as enterorrhagia, whereas hematochezia is more associated with neoplasms, hemorrhoidal disease, and angiodysplasia. However, no significant differences were found between the groups regarding sex, ethnicity, or comorbidities.



Conflict of Interest

The authors report no conflict of interest.

Authors' Contributions

CHMS: conceived the study, critically revised the manuscript; CVC: collected, analyzed the data, drafted the first version of the manuscript; BEZ: contributed to data collection, interpretation; COSR: contributed to data collection, interpretation; OAF: contributed to data collection, interpretation.



Address for correspondence

Camila Vieira Chiquetti
Rua Petrópolis, 1220, Block 5, Apartment 103, Campo Grande – MS
Brazil   

Publication History

Received: 29 April 2025

Accepted: 16 July 2025

Article published online:
25 September 2025

© 2025. The Author(s). This is an open access article published by Thieme under the terms of the Creative Commons Attribution 4.0 International License, permitting copying and reproduction so long as the original work is given appropriate credit (https://creativecommons.org/licenses/by/4.0/)

Thieme Revinter Publicações Ltda.
Rua Rego Freitas, 175, loja 1, República, São Paulo, SP, CEP 01220-010, Brazil

Bibliographical Record
Camila Vieira Chiquetti, Carlos Henrique Marques dos Santos, Bárbara Érnica Zeferino, Carlos Otávio da Silva Ribeiro, Odinilson Almeida Fonseca. Comparative Analysis of Lower Gastrointestinal Bleeding Due to Diverticular Disease Versus Other Etiologies. Journal of Coloproctology 2025; 45: s00451810612.
DOI: 10.1055/s-0045-1810612

Zoom
Fig. 1 Etiologies of Lower Gastrointestinal Bleeding (LGIB). The results are presented in relative frequency. *Others: colostomy varices; ileofemoral portosystemic shunt; Crohn's disease; fissure; intestinal pseudolipomatosis.
Zoom
Fig. 2 Chart illustrating the percentage of patients according to etiology and the type of treatment received. Each column represents the percentage value. Each column represents the percentage value. *Significant difference compared to patients with HD, angiodysplasia, and neoplasia for the same type of treatment (chi-square test with Bonferroni correction, p < 0.05).