Endoscopy 2014; 46(06): 539
DOI: 10.1055/s-0034-1365407
Letters to the editor
© Georg Thieme Verlag KG Stuttgart · New York

Reply to Yamashita et al.

Cesare Hassan
Angelo Zullo
Allesandro Repici
Further Information

Publication History

Publication Date:
28 May 2014 (online)

We read with great interest the letter from Yamashita et al. reporting a statistically significant association between antithrombotic therapy and the risk of biopsy bleeding after cold biopsy in the stomach [1]. This finding is in line with recent experience from our group after cold polypectomy for subcentimetric colorectal polyps, where bleeding risk was higher in patients on antiplatelet therapy (6.2 % vs. 1.4 %; P < 0.001) [2]. The main difference, however, is that in our lower gastrointestinal experience all the post cold polypectomy bleedings were mild and immediate, and therefore successfully treated with hemostatic clips, while in the series of Yamashita et al. both of the bleeding episodes were severe and delayed, incurring additional costs because of hospitalization and the need for repeat endoscopy [1].

Despite the severity of both the bleeding episodes described [1], it is unclear whether it is really necessary to interrupt antithrombotic treatment before cold biopsy in the upper gastrointestinal tract. First, the absolute risk of bleeding is really small, so that antithrombotic therapy would need to be interrupted in 250 cases to prevent, in the best scenario, one case of bleeding [1]. Second, it cannot be excluded that such a risk reduction may be offset by a higher risk of cardiovascular accident directly due to the biopsy-related interruption of antithrombotic therapy. In addition, it is unclear whether the preventive interruption of antithrombotic therapy would really reduce the risk of post-biopsy bleeding, since large randomized trials are lacking. Third, it should be noted that the delayed bleeding episodes occurred in one elderly patient with continuing dual antiplatelet therapy (aspirin and ticlopidine), and in one elderly patient in whom cilostazol treatment – not a typical antithrombotic therapy – had actually been stopped 7 days before endoscopy [1]. This would suggest that the risk of bleeding is present only in those patients with continuing dual antiplatelet therapy, whilst the authors’ data would not support the idea of a higher risk – if any at all – in those taking a single antithrombotic therapy [1]. Fourth, in the experience of Yamashita et al., the adoption of a policy of prudence was not apparently associated with a significant drop in the post-biopsy bleeding risk in patients taking antithrombotic agents [1]. Fifth, the observation that no bleeding episodes occurred in several cirrhotic patients enrolled in different studies, despite the presence of both clotting impairment and thrombocytopenia in those patients, further suggests the safety of performing biopsies on gastric mucosa [3].

A possible compromise between the need to sample upper and lower gastrointestinal mucosa and a slightly increased risk of bleeding related to antithrombotic therapy could be the exploitation of in vivo endoscopic prediction of lesion histotype, based on advanced endoscopic imaging techniques, such as electronic chromoendoscopy or magnification endoscopy [4] [5]. In particular, for the detection of gastric precancerous lesions, advanced endoscopic imaging has shown promising accuracy and interobserver agreement in a large multicenter study [6]. This may be even more true where there is suspicion of cancer, as in the two cases of bleeding reported by Yamashita et al. [1]. Thus, the higher risk of bleeding related to cold biopsy or polypectomy in patients on antithrombotic treatment should warrant a pre-biopsy endoscopic stratification of the neoplastic risk; thus only those patients who might actually gain a clinically relevant benefit from either mucosal sampling or polypectomy would be exposed to the risk of bleeding.