Abstract
Endoscopic submucosal dissection (ESD) is almost always performed with a sedative because of the longer procedure times involved. The risk of post-ESD deep vein thrombosis (DVT) has been reported as relatively high, and D-dimer levels are sometimes elevated after ESD. This retrospective study evaluated factors affecting changes in D-dimer levels from before to after ESD to identify causes of elevated D-dimer levels after ESD. This retrospective analysis included 117 patients with gastrointestinal tumors resected using ESD. After excluding eight patients with pre-ESD levels of D-dimer >1.5 μg/mL, factors correlating with changes in D-dimer from before to after ESD were analyzed using logistic regression analysis in 109 patients. Sedation was accomplished primarily using midazolam, but, because the sedative effect of midazolam shows marked inter-individual variability, a "corrected midazolam dose" was determined by dividing the total midazolam dose by the initial dose to correct for inter-individual differences in the sedative effect of midazolam. This value was used as one potential explanatory variable in the subgroup analysis of the 103 patients who received midazolam. In the subgroup analysis using the corrected midazolam dose as an explanatory variable, only the corrected midazolam dose correlated with a change in D-dimer ≥1.0 μg/mL in multivariate analysis (odds ratio (OR) = 1.5, 95% confidence interval (CI) 0.43-0.95; p = 0.030). The corrected midazolam dose correlated with increases in post-ESD D-dimer levels. This potential relationship indicates that patients undergoing ESD and requiring extended sedation may be at increased risk of DVT.