Effectiveness of suction valve button removal in retrieving resected colon polyps for better histological assessment: Propensity score matching analysis
Abstract
Background and Aim
Fragmentation of endoscopically resected colorectal polyps during retrieval is one of the limitations for appropriate pathological diagnosis; however, little is known about steps to reduce it. We aimed to evaluate the effect of removing the suction valve button, which is one of the intricate parts of the endoscope, during polyp suction retrieval for fragmentation and pathological diagnosis.
Methods
We retrospectively reviewed the polyps retrieved by suctioning. We used the propensity score matching method to adjust for difference between the button-attached and button-removed groups. Outcomes of fragmentation, pathological diagnosis of non-neoplastic polypectomy (NNP), and pathological cut-end among neoplastic lesions were evaluated.
Results
On the basis of propensity score matching, 322 pairs of cases were selected for analysis. After matching, the difference in the variables between the two groups was closely balanced. The fragmentation rate was significantly different between the groups (button-attached 36.6% vs. button-removed 22.4%, P < 0.001). No significant difference in the NNP rate was observed between the two groups (button-attached 8.4% vs. button-removed 9.9%). The rate of unclear cut-end was lower in the button-removed group than in the button-attached group (16.2% vs. 9.1%, P = 0.010). In the multivariate analysis, button removal was independently associated with reductions of fragmentation (odds ratio [OR] 0.508, P < 0.001) and unclear cut-end (OR 0.503, P = 0.009). Also, NNP was associated with smaller lesion size (OR 0.364, P < 0.001), but not with button removal.
Conclusions
Suction valve button removal during polyp suction retrieval was an effective method to decrease the rates of fragmentation and histological unclear margins.
Conflict of Interest
Author N.K. is an Associate Editor of Digestive Endoscopy. Other authors declare no Conflict of Interests for this article.
