Comparison of postpolypectomy bleeding between epinephrine and saline submucosal injection for large colon polyps by conventional polypectomy: A prospective randomized, multicenter study

Table 1 Baseline characteristics of the studied patients

	Epinephrine group(n = 244)	Normal saline group(n = 242)	P
Clinical features
Age (mean ± SD)	56.0 ± 11.4	56.8 ± 11.3	0.429
Sex (Male%)	195 (70.9%)	184 (64.3%)	0.096
BMI (mean ± SD)	23.0 ± 2.9	23.8 ± 2.7	0.572
Endoscopic features
Total No. of polyps	275	286	0.587
Mean No. of polyps	1.13 ± 1.6	1.18 ± 2.0	0.149
Mean size of polyps (mm)	14.5 ± 5.7	15.0 ± 6.8	0.318
Macroscopic form of polyps			0.427
Pedunculated (Ip)	62 (22.5%)	66 (23.1%)
Semipedunculated (Isp)	93 (33.8%)	107 (37.4%)
Sessile (Is)	70 (25.5%)	80 (28.0%)
Superficial elevated type (IIa)	50 (18.1%)	33 (11.5%)
Distribution of polyps			0.809
Left colon (descending-)	131 (53.7%)	126 (52.1%)
Right colon (-transverse)	51 (20.9%)	48 (19.8%)
Both	62 (25.4%)	68 (28.1%)
Pathologic features
Non-neoplastic	23 (8.4%)	28 (9.8%)	0.557
Neoplastic
Dysplasia	184 (66.9%)	178 (62.2%)	0.269
HG dysplasia + CIS1	44 (16.0%)	51 (17.8%)	0.626
Cancer	24 (8.7%)	29 (10.2%)	0.557
Procedure related features
Inexpert operator2	42 (17.2%)	48 (19.8%)	0.626
En bloc resection	261 (94.9%)	268 (93.7%)	0.070

¹HG + CIS, high grade dysplasia + carcinoma in situ.

²Inexpert operator, experience in polypectomy < 2 years.

Table 2 Postpolypectomy bleeding, n (%)

	Epinephrine group	Normal saline group	P value
	(n = 244)	(n = 242)
Early bleeding	11 (4.5)	21 (8.7)	0.065
Late bleeding	1 (0.4)	4 (1.7)	0.154

Table 3 Risk factors for postpolypectomy bleeding (multivariable analysis)

Risk factor	Adjusted odds ratio(95% CI)1	P
Size (> 2 cm)	1.07 (1.01, 1.14)	0.034
Injection (normal saline only)	1.31 (0.57, 2.99)	0.527
Pathology (neoplastic)	9.88 (1.26, 77.75)	0.029
Morphology (protruding type: Ip, Isp, Is)	0.59 (0.18, 1.95)	0.393
Inexpert operator (< 2 yr)	2.62 (0.98, 7.06)	0.056
Hospitalization (not hospitalized)	2.14 (0.57, 8.04)	0.261

¹Adjusted odds ratios are calculated from a multivariate logistic regression model except for age, sex and procedure time.