The incidence of colonoscopy has risen in response to the growing interest in colorectal cancer (CRC) prevention.^1,2 Polypectomy, a key technique for preventing CRC, is known to reduce the morbidity rate of colon cancer by nearly 50%. However, complications such as bleeding, perforation, post-polypectomy coagulation syndrome, and stricture formation can arise after the procedure.³

Bleeding is the most common complication associated with polypectomy, followed by perforation.³ The incidence of immediate post-polypectomy bleeding (IPPB) is 2.8%, whereas delayed post-polypectomy bleeding (DPPB) occurs in 0.2% to 2.2% of cases.^4–6 Common hemostatic methods for managing post-polypectomy bleeding (PPB) include epinephrine injection at a dilution of 1:10,000, non-contact thermal methods such as argon plasma coagulation (APC), contact thermal methods including hemostatic coagulation forceps and heat probes, and mechanical methods like the application of hemoclips, over-the-scope clips, and detachable loops.

Most PPB can be controlled with clip placement, with or without epinephrine injection. Although intraprocedural bleeding is relatively easy to control by promptly identifying and addressing the source, delayed identification and hemostasis in DPPB can lead to serious complications for the patient. Furthermore, if not quickly and effectively managed, severe bleeding may result in significant blood loss, hypovolemic shock, organ failure, and potentially death. Therefore, when severe bleeding is suspected, hospitalization is necessary for close monitoring, identification of the bleeding source, and hemostatic interventions. Such hospitalization can lead to economic losses and reduce the cost-effectiveness of the procedure. To improve procedural safety and predict PPB, researchers have endeavored to identify associated risk factors.^7–9 These risk factors for DPPB include hypertension and cardiovascular disease, polyps located in the right colon, polyps larger than 10 mm, and endoscopists with less than 1 year of experience or fewer than 300 polypectomy cases.⁷ The resumption of anticoagulation therapy after polypectomy and larger polyp diameter have also been strongly associated with an increased risk of DPPB.⁸ Separately, the size and number of resected polyps, as well as the endoscopist’s experience, have been implicated as independent risk factors.⁹

Nevertheless, the clinical outcomes of endoscopic hemostasis for DPPB and the risk factors for rebleeding after such intervention are not well understood. This study aims to evaluate the clinical outcomes and effectiveness of endoscopic hemostasis in treating DPPB, as well as to identify the risk factors associated with rebleeding following the intervention.

Patients

Overall, 289 patients who developed DPPB and underwent endoscopic hemostasis were retrospectively identified from five university hospitals in Korea between January 2005 and May 2018. The study included patients who had colon polyps larger than 5 mm, underwent colon polypectomy, and developed PPB between 24 hours and 3 weeks after the procedure. The exclusion criteria ruled out patients with colon polyps smaller than 5 mm, individuals experiencing PPB within 24 hours or after 3 weeks, and cases of hematochezia or melena not associated with polypectomy. Cases were also excluded if they lacked definitive evidence of bleeding on endoscopy or had incomplete medical records.

Terms and data analysis

PPB can be categorized into two types: IPPB, which occurs within 24 hours, and DPPB, which occurs after 24 hours. Our analysis of DPPB was based on patient characteristics, lesion features, and technical factors. Patient characteristics included age, sex, and comorbidities such as diabetes mellitus, hypertension, chronic kidney disease, and coronary artery disease, as well as the use of antiplatelet or anticoagulant medications. Lesion features encompassed polyp morphology, size, location, and pathological diagnosis. Technical factors involved polypectomy techniques (cold forceps, cold snare, and hot snare), hemostatic methods (epinephrine injection, APC, and the application of hemoclips), and sedation status. We also evaluated the rate of rebleeding in relation to the duration of the procedure and the length of hospitalization.

Hemostatic procedures

Patients with suspected PPB underwent re-examination with a standard flexible colonoscope (CF-H260AL/CF-H260AI/CF-H290ECI; Olympus Corp.), used with or without a transparent cap. The endoscopist decided whether to use sedation based on the patient’s condition, their preferences, and the setting in which the emergency endoscopy took place. If bleeding was detected, hemostasis was achieved using one or more of the following methods: hemoclips, epinephrine injection, APC, or hemostatic coagulation forceps. For procedures involving APC or hemostatic coagulation forceps, an electrosurgical unit (VIO 300D; Erbe Elektromedizin GmbH) was utilized.

Ethical statement and statistical analysis

This study was conducted in accordance with the Declaration of Helsinki. All sample collections and clinical data recordings were approved by the relevant institutional ethics committee (SCHCA 2018-07-057).

Unadjusted differences in rebleeding rates following endoscopic hemostasis for DPPB across study variables were compared using the chi-square test. To identify risk factors for rebleeding after endoscopic hemostasis for DPPB, multiple logistic regression analysis was utilized to calculate odds ratios (ORs) and 95% confidence intervals (CIs). The risk factors analyzed included patient characteristics (age, sex, and comorbidities), polyp attributes (morphology, size, location, and histological findings), polypectomy techniques, and characteristics of the bleeding, with adjustments made for patient age and sex. A P-value < 0.05 was considered to indicate statistical significance. All statistical analyses were performed using STATA (version 14.0; Stata Corp.).

Endoscopic hemostasis was successfully performed in all 289 cases of DPPB. The endoscopic techniques employed for hemostasis included hemoclipping in 87.9% of the cases, epinephrine injection in 24.9%, APC in 18.0%, and coagulation with hemostatic forceps in 10.7%. Regarding the number of methods used for hemostasis, 65.7% of the cases required a single method, 27.0% utilized two methods, and 7.35% required three methods. The mean total time for the endoscopic procedure was 9.7 ± 8.7 minutes, the mean number of hemoclips used per case was 3.0 ± 2.0, and the average hospital stay was 3.2 ± 2.6 days.

Rebleeding occurred in 5.2% (15/289) of cases following endoscopic hemostasis for DPPB (Fig. 1). No significant differences were found between the rebleeding and non-rebleeding groups regarding patient characteristics such as age, sex, comorbidities (hypertension, diabetes mellitus, and chronic kidney disease), or the use of antithrombotic agents (Table 1). Rebleeding after endoscopic hemostasis was significantly associated with polyp size (< 10 mm: 2.8%, 10 mm–19 mm: 5.6%, ≥ 20 mm: 13.5%, P = 0.030) and sedation status (sedated: 1.8%, not sedated: 7.3%, P = 0.040). However, polyp morphology, location, and pathological findings were not significantly related to rebleeding. Additionally, the technique used for polypectomy (cold biopsy, cold snare, and hot snare with or without injection) (P = 0.786), features of the bleeding (pulsatile, oozing, or exposed vessels without bleeding) (P = 0.298) (Table 2), endoscopic hemostasis techniques, and number of hemostasis methods applied showed no significant association with rebleeding. Multivariate analysis revealed that polyp size (OR, 5.02; 95% CI, 1.25–20.13; P = 0.023) was significantly associated with rebleeding after endoscopic hemostasis for DPPB. In all 15 cases of rebleeding, a second endoscopic hemostasis was successfully performed without the need for embolization or surgery. Furthermore, no additional bleeding or perforations occurred during either the first or the second hemostasis (Tables 3, 4).

Table 1 . Demographic Characteristics of Patients.

Characteristic	Total (n = 289)	Rebleeding		P-value
		Negative (n = 274)	Positive (n = 15)
Age (yr)	55.8 ± 13.2	55.8 ± 13.3	55.8 ± 13.3	0.619
Sex				0.431
Male	227 (78.5)	214 (94.3)	13 (5.7)
Female	62 (21.5)	60 (96.8)	2 (3.2)
Comorbidity
Hypertension	104 (36.0)	100 (96.2)	4 (3.8)	0.440
Diabetes mellitus	36 (12.5)	34 (94.4)	2 (5.6)	0.916
Chronic kidney disease	15 (5.2)	15 (100)	0 (0.0)	0.352
Coronary artery disease	27 (9.3)	26 (96.3)	1 (3.7)	0.715
Antiplatelet or anticoagulation agent				0.820
Administered	55 (19.0)	53 (96.4)	2 (3.6)
Not administered	234 (81.0)	221 (94.4)	13 (5.6)

Values are presented as mean ± standard deviation or number (%).

Table 2 . Comparison of Endoscopic and Histological Features Between Groups.

Characteristic	Total (n = 289)	Rebleeding		P-value
		Negative (n = 274)	Positive (n = 15)
Polyp morphology				0.640
Pedunculated	62 (21.5)	59 (95.2)	3 (4.8)
Sessile	176 (60.9)	168 (95.5)	8 (4.5)
Flat	51 (17.6)	47 (92.2)	4 (7.8)
Polyp size (mm)				0.030
< 10	145 (50.2)	141 (97.2)	4 (2.8)
10–19	107 (37.0)	101 (94.4)	6 (5.6)
≥ 20	37 (12.8)	32 (86.5)	5 (13.5)
Polyp location				0.670
Right (ascending to transverse)	167 (57.8)	160 (95.8)	7 (4.2)
Left (descending to sigmoid)	76 (26.3)	71 (93.4)	5 (6.6)
Rectum	46 (15.9)	43 (93.5)	3 (6.5)
Pathologic diagnosis				0.141
Hyperplastic polyp	59 (20.4)	59 (100)	0 (0.0)
Adenoma	214 (74.0)	201 (93.9)	13 (6.1)
Cancer	16 (5.6)	14 (85.0)	2 (15.0)
Sedation				0.040
Yes	111 (38.4)	109 (98.2)	2 (1.8)
No	178 (61.6)	165 (92.7)	13 (7.3)
Polypectomy technique				0.786
Cold biopsy polypectomy	10 (3.5)	10 (100)	0 (0.0)
Cold snare polypectomy	6 (2.1)	6 (100)	0 (0.0)
Hot snare polypectomy without injection	13 (4.5)	12 (92.3)	1 (7.7)
Endoscopic mucosal resection	260 (89.9)	246 (94.6)	14 (5.4)
Bleeding feature				0.298
Pulsatile	10 (3.5)	9 (90.0)	1 (10.0)
Oozing	169 (58.5)	158 (93.5)	11 (6.5)
Exposed vessel without bleeding	110 (38.0)	107 (97.3)	3 (2.7)

Values are presented as number (%).

Table 3 . Comparison of Performance of Endoscopic Hemostasis Between Groups.

Procedural feature	Total (n = 289)	Rebleeding		P-value
		Negative (n = 274)	Positive (n = 15)
Method(s) of endoscopic hemostasis
Epinephrine injection	72 (24.9)	68 (94.5)	4 (5.5)	0.872
Argon plasma coagulation	52 (18.0)	48 (92.3)	4 (7.7)	0.369
Hemostatic coagulation forceps	31 (10.7)	31 (100)	0 (0.0)	0.168
Hemoclips	254 (87.9)	240 (94.5)	14 (5.5)	0.507
Mean number of hemoclips used per case	3.03 ± 1.95	3.03 ± 1.96	3.03 ± 1.81	0.841
Total number of types of hemostatic procedures used				0.851
One method	190 (65.7)	181 (95.3)	9 (4.7)
Two methods	78 (27.0)	73 (93.6)	5 (6.4)
Three methods	21 (7.3)	20 (95.2)	1 (4.8)
Method of secondary hemostasis
Endoscopic hemostasis	15 (100)	0 (0.0)	15 (100)
Embolization	0 (0.0)	0 (0.0)	0 (0.0)
Surgery	0 (0.0)	0 (0.0)	0 (0.0)
Perforation	0 (0.0)	0 (0.0)	0 (0.0)
Mean procedure time (sec)	580.4 ± 523.6	574.2 ± 509.2	694.1 ± 755.3	0.389
Mean hospital stay (day)	3.2 ± 2.6	3.1 ± 2.2	5.7 ± 5.9	< 0.001

Values are presented as number (%) or mean ± standard deviation.

Table 4 . Multivariate Analysis of Risk Factors for Rebleeding Following Endoscopic Hemostasis for Delayed Post-Polypectomy Bleeding.

Characteristic	Odds ratio	95% confidence interval	P-value
Age (yr)	0.99	0.95–1.04	0.748
Sex	1.56	0.33–7.45	0.575
Sedation	0.27	0.06–1.27	0.098
Polyp size (mm)	5.02	1.25–20.13	0.023

Adjusted for age and sex.

Figure 1. Rebleeding rates over time following hemostasis for delayed post-polypectomy bleeding (DPPB).

Colonoscopic polypectomy is a key technique for the removal of colorectal neoplasms, and research has demonstrated that it leads to a significant 53% reduction in mortality associated with CRC.¹⁰ However, complications can arise, including PPB. A meta-analysis of 21 studies examining complications from colonoscopy reported a PPB incidence of 0.98%, with a decreasing trend observed from 2001 to 2010.¹¹ Clinically, relative to IPPB, DPPB tends to be more serious due to the greater volume of blood loss, and the initiation of appropriate treatment may be delayed.¹² Consequently, DPPB can constitute a significant complication, often requiring hospitalization, blood transfusion, and repeat endoscopic intervention.

In efforts to prevent or predict DPPB, research has been conducted to identify its risk factors.^13–15 A study of 3,788 patients revealed that greater polyp size, less endoscopist experience, and polyp location in the right hemicolon were key risk factors for DPPB.¹³ In an analysis of 16,925 patients, young age, male sex, hypertension, polyps measuring 1 cm or larger, location in the proximal colon, and type of operative technique employed were identified as significant DPPB risk factors.¹⁴ Additionally, a meta-analysis encompassing 12 articles and 14,313 patients found that cardiovascular disease (OR, 1.55), hypertension (OR, 1.53), polyp size exceeding 10 mm (OR, 3.41), and polyp positioning in the right colon (OR, 1.60) represented significant risk factors for DPPB.¹⁵

When DPPB occurs or is suspected, endoscopic therapy should be provided to patients with high-risk endoscopic stigmata of bleeding, such as active bleeding (spurting or oozing), a non-bleeding visible vessel, or an adherent clot.¹⁶ Guidelines recommend managing PPB with mechanical therapy (such as clips) or contact thermal therapy, with or without the adjunctive use of diluted epinephrine injection. The use of epinephrine injection alone or APC is not recommended.^16,17 In this study, hemoclips were the most frequently used method for achieving initial hemostasis after DPPB, utilized in 87.9% of cases. The mean number of hemoclips applied per case was 3.03 ± 1.95. A single hemostasis method was employed in 65.7% of cases, two methods in 27.0%, and three methods in 7.3%. Initial endoscopic hemostasis was successful in all cases of DPPB. Although DPPB can typically be managed with endoscopic hemostasis techniques, rebleeding may occur in approximately 4.2% to 9.5% of cases after successful endoscopic hemostasis.^14,18–20 Rebleeding is burdensome to patients, as the associated hospitalization increases medical costs, and repeated colonoscopies can be painful. Moreover, the need for a second endoscopic hemostasis procedure elevates patient discomfort and healthcare costs. In this study, while initial endoscopic hemostasis for DPPB was successful in all cases, the rate of rebleeding necessitating a second hemostasis was 5.2%, aligning with findings from previous studies.

The risk factors for rebleeding following successful endoscopic hemostasis for DPPB have not been clearly understood. A retrospective study of 198 patients with DPPB revealed that 15 (7.6%) experienced rebleeding. Although the mean polyp size was greater in the rebleeding group, this difference was not statistically significant (18.0 mm ± 18.5 mm vs. 12.2 mm ± 13.3 mm; P = 0.217). The study also indicated that the use of a greater number of clips and the combination of clips with other hemostatic methods were associated with increased risk of rebleeding after endoscopic hemostasis for DPPB.²⁰ In the present study, as polyp size increased, the rebleeding rate increased significantly (OR, 5.02; 95% CI, 1.25–20.13; P = 0.023). Additionally, the use of sedation was significantly linked to a reduced risk of rebleeding (sedated: 1.8%, not sedated: 7.3%, P = 0.040) after endoscopic hemostasis for DPPB. However, neither endoscopic hemostasis technique nor the number of hemostasis methods used displayed a significant relationship with rebleeding. Risk factors for DPPB, such as anticoagulation, comorbidities, and polyp location, do not seem linked to rebleeding after initial hemostasis of DPPB. However, polyp size represents the singular risk factor for both DPPB and rebleeding post-hemostasis, warranting particular attention.

This study had several limitations. Although we collected data retrospectively from multiple centers, the low incidence of DPPB and the even lower rate of rebleeding constrained our ability to amass many cases with rebleeding after DPPB. The relatively small sample size of patients who experienced rebleeding following hemostasis for DPPB limited the generalizability of our findings. Given the retrospective, multicenter nature of the study, it was challenging to account for variations in the experience and proficiency of endoscopists. Furthermore, some cases of rebleeding may have been subclinical or not documented in medical records, potentially leading to an underestimation of the actual rebleeding rate. Lastly, the study was conducted at high-level medical facilities; thus, the results may not fully represent polypectomy procedures performed in smaller hospitals or outpatient settings.

Rebleeding following endoscopic hemostasis for DPPB can have devastating consequences for patients, impacting not only medical costs but also the patient-physician relationship. Therefore, minimizing such occurrences is crucial. In this study, larger polyp size was associated with a higher risk of DPPB and rebleeding after hemostasis, suggesting that more careful endoscopic management is necessary for these cases. Furthermore, the use of sedation, when feasible, may alleviate patient discomfort and facilitate more thorough hemostasis. To improve the statistical power and reliability of these findings, future research would benefit from a larger cohort.

In conclusion, polyp size and sedation status appear to be associated with rebleeding after endoscopic hemostasis for DPPB. Endoscopic hemostasis in DPPB seems to represent a safe and effective intervention.

None.

Research data are not shared.

No potential conflict of interest relevant to this article was reported.