Endoscopic mucosal resection (EMR) and endoscopic submucosal dissection (ESD) are curative techniques for gastric neoplasm; ESD has advantages over EMR regarding en bloc resection of large-sized lesions and submucosal fibrosis [1-4].

Although ESD is widely considered a safe technique, complications including bleeding and perforation remain major concerns. While coagulation therapy after ESD and the use of proton pump inhibitor (PPI) premedication have gradually decreased the frequency of delayed bleeding, delayed bleeding still occurs in about 5% of cases [5-7]. Therefore, it is clinically important to prevent delayed bleeding.

Endoscopic findings of peptic ulcer bleeding, including active or recent hemorrhage, are classified by using the Forrest classification [8], a well-known and useful tool for evaluating and predicting the risk of recurrent bleeding following endoscopic hemostasis [9]. Peptic ulcers with a Forrest classification from Ia (spurting arterial vessel) to IIb (adherent clot) are more likely to rebleed after initial hemostasis; endoscopic hemostasis during the first endoscopy is recommended to prevent rebleeding from these lesions [10,11], In addition, second-look endoscopy (SLE) is reported as useful in preventing rebleeding in patients with high-risk peptic ulcer bleeding [12,13], Therefore, to evaluate post-ESD bleeding and achieve hemostasis of high-risk lesions, SLE is commonly performed in many hospitals after ESD. However, the use of SLE after ESD is still controversial.

Although a recent retrospective study reported the usefulness of SLE for preventing post-ESD bleeding on the basis of less frequent delayed bleeding after SLE (one case in 432 patients vs. eight cases in 440 patients) [14], other studies have reported the non-significant effect of SLE on clinical outcomes including delayed bleeding [5,15,16], Furthermore, one prospective study suggested that prophylactic coagulation for post-ESD ulcers with a high-risk Forrest classification may have minor effects on preventing delayed bleeding [17].

One major difference between peptic ulcers and post-ESD ulcers is the conditions associated with their formation [18]. Peptic ulcers are usually associated with low gastric pH; rebleeding is more likely to occur in patients with these ulcers. Therefore, SLE for peptic ulcer bleeding is considered useful. PPI premedication relatively raises the pH of the post-ESD ulcers’ environment, and heals post-ESD ulcers faster than peptic ulcers. However, the risk of rebleeding may be decreased, irrespective of the use of preventive hemostasis on post-ESD ulcers [17-19]

Despite this controversy, it is difficult to recommend discontinuing SLE and many endoscopists continue to perform SLE in their clinics. What valuable information, if any, can we obtain from SLE? Can we predict delayed bleeding by using SLE? As described in this issue of Clinical Endoscopy, Kim et al. [20] retrospectively analyzed 581 lesions to study delayed bleeding following SLE. To predict delayed post-ESD bleeding, they focused on the usefulness of the Forrest classification of post-ESD ulcers during SLE. To investigate recent hemorrhage or potential bleeding, each SLE was performed one day after each ESD; the post-ESD ulcers were classified according to the Forrest classification. Using these post-ESD ulcer classifications, they then determined bleeding rates after SLE. When post-ESD ulcers with high-risk Forrest classifications were observed, the endoscopists made subjective decisions regarding performing prophylactic hemostasis. Delayed post-ESD bleeding occurred in 42 of the 581 patients (7.2%). During SLE, Forrest Ib, IIa, and IIb ulcers were identified in 8.2%, 8.0%, and 13.7% of cases, respectively. All Forrest Ia lesions were diagnosed by emergency bleeding occurring within 24 hours of performing ESD, and no Forrest Ia lesions were found during SLE. Multivariate analyses determined that a specimen size ≥40 mm (odds ratio [OR], 3.03; p=0.003), and a high-risk Forrest classification (Forrest Ib+IIa+IIb; OR, 6.88; p<0.001) were risk factors for delayed post-ESD bleeding. In accordance with a previous study [17], the rate of delayed post-ESD bleeding was not significantly different between the prophylactic hemostasis and non-hemostasis groups.

Based on this study, Forrest classification of post-ESD ulcers may be useful in predicating a high risk of delayed post-ESD bleeding, but routine prophylactic hemostasis for post-ESD ulcers with a high-risk Forrest classification during SLE is not useful in preventing delayed post-ESD bleeding. Predicting a high risk of delayed post-ESD bleeding may help physicians establish therapeutic plans, influence decision timing for oral intake/discharge, and assist them in efforts to educate their patients on taking regular medications such as PPIs, or on visiting the hospital as soon as possible if associated symptoms develop. However, this doesn’t mean that SLE should be done in all patients. Therefore, future prospective studies are needed to define subjects with a high risk of delayed post-ESD bleeding, in whom SLE is needed.