Effectiveness and safety of postoperative hemostatic drugs in colorectal cancer: A retrospective analysis of a prospectively collected multicenter database using inverse probability of treatment weighting.

BACKGROUND: Hemostatic drugs are frequently prescribed after colorectal cancer (CRC) surgery to prevent bleeding. However, in real-world practice, their net clinical benefit and thrombotic safety remain uncertain. This study aimed to evaluate the effectiveness and safety of prophylactic postoperative hemostatic drug use after CRC surgery. METHODS: This retrospective analysis was based on a prospectively collected multicenter CRC-venous thromboembolism (VTE) database involving 46 tertiary centers across 17 provinces in China between May 2021 and May 2022. Adults who underwent elective curative CRC surgery were included and followed up for 28 days. Exposure was defined as initiation of any prophylactic postoperative hemostatic drug (tranexamic acid, hemocoagulase, or carbazochrome sodium sulfonate) within 0-48 h after surgery in the absence of active bleeding at the first administration. Primary outcome was postoperative bleeding within 28 days; a surrogate outcome was hemoglobin decrease (dHb) ≥20 g/L. The secondary outcome was imaging-confirmed VTE, defined as postoperative DVT and/or PE; DVT was systematically screened within prespecified follow-up windows, whereas PE was assessed when clinically suspected. Inverse probability of treatment weighting (IPTW) based on propensity scores was used to balance baseline covariates, followed by weighted logistic regression. RESULTS: Among 1810 patients, 322 (17.8%) received prophylactic postoperative hemostatic drugs. After IPTW, all covariates achieved adequate balance, with standardized mean differences <0.10. Hemostatic drug use was not associated with reduced postoperative bleeding (2.9% vs. 1.7%; weighted odds ratio [OR]: 1.72, 95% confidence interval [CI]: 0.77-3.86) or dHb ≥20 g/L (21.4% vs. 26.1%; weighted OR: 0.77, 95% CI: 0.55-1.06). In contrast, hemostatic drug use was associated with a higher postoperative VTE rate (13.8% vs. 9.5%; weighted OR: 1.51, 95% CI: 1.02-2.24), mainly driven by DVT without PE, whereas PE events were rare and did not differ significantly between groups. Additional analyses showed that this association was also observed for symptomatic VTE (4.7% vs. 2.3%; weighted OR: 2.12, 95% CI: 1.03-4.37), whereas the association for asymptomatic/screening-detected VTE was not statistically significant after IPTW. Most VTE events with a DVT component involved distal DVT. Sensitivity analyses using winsorized weights yielded consistent estimates, and Propensity score trimming excluded no participants, supporting adequate overlap. In multivariable analysis, tumor-node-metastasis stage III-IV (OR: 3.20, 95% CI: 1.31-7.79) and intraoperative blood loss ≥50 mL (OR: 3.15, 95% CI: 1.20-8.22) were independent predictors of postoperative bleeding. CONCLUSIONS: In this retrospective analysis of a prospectively collected multicenter database, routine prophylactic postoperative hemostatic drug use after CRC surgery was not associated with a lower bleeding risk, but was associated with a higher postoperative VTE rate. Given the limited number of symptomatic events, the predominance of distal DVT, and the potential for residual confounding, this finding should be interpreted as a hypothesis-generating thrombotic safety signal rather than definitive causal evidence. Prophylactic hemostatic drugs should be used individually, with careful consideration of thrombotic risk.