UI-EWD powder is a newly developed topical hemostatic agent. Use of this powder in refractory upper GI bleeding, non-variceal bleeding, and diffuse tumor bleeding has been described as safe and efficient for the prevention of rebleeding [18–21]. In the present study, we showed another promising result of UI-EWD on LGIB in that it prevented cumulative re-bleeding rate down to 5.5%, showing a greater advantage in sites that are difficult to access by endoscopy or those that are larger with diffuse oozing bleeding.
The rebleeding rate in the UI-EWD group was 5.5%, which is a considerable improvement compared to the rate of 12.8% observed with previously reported hemostatic powders in LGIB [16]. In addition to the common advantages of hemostatic powders, including easier accessibility and greater applicability to larger and diffuse oozing bleeding lesions [24, 25], differences in the composition of hemostatic powders might affect the hemostatic efficiency. Although there has been no head-to-head comparison, UI-EWD contains aldehyde dextran as a main component mixed with ε-poly amino acid at a 4:1 ratio, while other commercially available powders are mainly composed of polysaccharides [20]. These differences in UI-EWD components are thought to lower the rebleeding rate by increasing the adhesion power from the Schiff base reaction through the cross-linking of amine and aldehyde dextran groups when they come into contact with moisture. In previous reports, the UI-EWD hydrogel was still present at 70.2% of the sprayed bleeding sites by second-look endoscopy at 24 h [20].
The characteristic of hemostatic powder working as locally, a shorter hemostatic effect is inevitable under the conditions of GI tract movement and food or feces passing. In addition, in clinical settings, it has been reported that the rebleeding rate further increases with time despite a high rate of successful initial hemostasis as a result of using hemostatic powder in upper and lower GI bleeding [26]. Moreover, there was a high rate of rebleeding after application of Hemospray (33.7%), with similar rates for both esophagogastroscopies and colonoscopies (35.6% vs. 23.1%). Most reported rebleeding occurred within 7 days (86.2%) [27]. However, in our report, the UI-EWD group did not show delayed rebleeding between 7 and 28 days after the initial successful hemostasis in LGIB. In our study, rebleeding was reported in three cases within 7 days; however, no rebleeding was reported after 7 days.
The medical records of the three patients with rebleeding were reviewed. The first patient showed rebleeding 1 day after endoscopic submucosal dissection of a 3-cm Yamada type III polyp diagnosed with rectal cancer. Owing to the taller (as opposed to wide) shape, there is a high possibility that a thick feeding vessel was hidden by the remnant submucosal injection, which can cause active arterial bleeding. The second patient showed rebleeding 4 days after initial hemostasis of a rectal ulcer with underlying liver cirrhosis and a low platelet count of 20,000. These high bleeding tendencies could not escape from the rebleeding. The last patient showed rebleeding 5 days after initial hemostasis of a transverse colon ulcer and took dual antiplatelet agents due to a history of cardiac stent insertion. In addition to the hemostasis failure of initial clipping on diffuse oozing bleeding ulcers, clipping itself might promote bleeding by GI movement. Although there were only three rebleeding cases, patients with a higher bleeding tendency or those who bled from the rectum. Further, a well-designed prospective study of various conditions that increase rebleeding is needed.
In the UI-EWD monotherapy group (Table 4), most of the applied cases were post-procedure or tumor bleeding (15/17, 88.2%), with sizes of bleeding site > 1 cm (16/17, 94.1%). In these cases, additional endoscopic hemostasis was hesitated because of the higher risk of delayed perforation. Despite these difficult circumstances, a single application of UI-EWD seemed to be sufficient to prevent rebleeding and allow initial hemostasis. Moreover, there was no reported perforation or colonic obstruction due to the single use of UI-EWD. This is meaningful in that applying mono-treatment of hemostatic powder is less traumatic and may be a reliable option for active oozing bleeding with larger size, especially for post procedure or tumor bleeding in LGIB. This is compatible with the findings of our previous report on the effectiveness of UI-EWD for upper GI tumor bleeding [18].
In the comparison of bleeding characteristics between the conventional therapy group and the UI-EWD therapy group, UI-EWD was more commonly applied to the hepatic flexure. This is known as one of the most difficult areas to access or to fix the endoscopic position due to its sharp angulation, and more frequently has larger sizes > 1 cm, even ≥ 4 cm. Even under these adverse conditions, the UI-EWD group showed a significantly lower cumulative rebleeding rate within 28 days (5.5%) than the conventional therapy group (17%).
Considering the adverse events of visceral perforation and splenic infarction reported previously when using Hemospray [28], we also evaluated UI-EWD-related adverse effects, including embolism, intestinal obstruction, and allergic reaction, due to the characteristics of the powder. However, there were no reported cases in the UI-EWD group compared to perforation or infection reported by coagulation treatment in the conventional therapy group. In case of perforation and embolism by Hemospray, it was considered due to the high pressure when injecting powder, which induces a patient’s risk of abdominal pain after applying Hemosrapy [16]. As a result of direct comparison of pressure between UI-EWD and Hemospray, UI-EWD was significantly lower at 7 psi than at 37 psi. These low injection pressures may have contributed to the safety of UI-EWD.
Our study has some limitations. First, it was a retrospective analysis at a single-center hospital, which may not be generalizable. Second, the patients in whom UI-EWD was applied were not randomly distributed and the decision to use is left to the discretion of the physicians, which could present a selection bias. Despite these limitations, bleeding sites in the UI-EWD therapy group were more difficult to endoscopically approach, such as hepatic flexure or sigmoid colon, and were more common in the longer diameter group. To obtain further information on the practical performance of UI-EWD in LGIB and to evaluate the feasibility and efficacy of LGIB, a well-designed, large-scale, prospective study is needed.
In conclusion, this is the first study to demonstrate the effectiveness and safety of UI-EWD in the treatment of LGIB. Application of UI-EWD resulted in a high rate of immediate hemostasis (100%) and a significant reduction in re-bleeding in patients with LGIB whose condition is not well controlled by conventional treatment (17% vs. 5%, p = 0.044). Based on our results, we suggest that the use of UI-EWD for hemostasis in LGIB is sufficiently effective and safe.