We assessed the quality of the included evidence using the GRADE (Grading of Recommendations, Assessment, Development, and Evaluations) framework. GRADE is a widely adopted, transparent system for evaluating evidence and formulating clinical recommendations. It classifies the certainty of evidence as high, moderate, low, or very low based on explicit criteria across five key domains: risk of bias, inconsistency, indirectness, imprecision, and publication bias. Applying these criteria, the overall evidence synthesized in this review was judged to be of moderate or high certainty.

Abdominoperineal resection (APR) for rectal cancer involves removing the rectum and anus and creating an end colostomy (10). The procedure is technically demanding due to the pelvic anatomy, with notable risks of presacral venous plexus hemorrhage and injury to adjacent structures such as the vagina, bladder, prostate, ureters, and pelvic nerves (11). Additionally, APR is associated with a high complication rate, as consistently reported in the literature. Specifically, for open APR, the risk of symptomatic VTE at 4 wk was 3.6%, and the risk of bleeding requiring transfusion was 21.5% (12). In contrast, for laparoscopic APR, the risk of symptomatic VTE was 1.1%, and the risk of bleeding requiring transfusion was 4.9% (12).

For rectal cancer sparing the anal sphincter, anterior resection, or abdominal proctosigmoidectomy, is the indicated surgical procedure (13). It entails the excision of the rectosigmoid segment and the subsequent anastomosis of the descending colon to the proximal rectum. This sphincter-preserving method, which precludes perineal dissection, thereby maintains intestinal continuity and optimizes postoperative functional outcomes. Sphincter-preserving surgery in the form of low anterior resection (LAR) is recognized as the gold standard for localized rectal cancer (14). LAR accounts for up to 80% of rectal cancer procedures (15). In 2023, a meta-analysis incorporating 18 clinical studies on anterior resection showed that for minimally invasive anterior resection, the risk of symptomatic VTE ranged from 0.8% to 3.2%, with a clinical relevant bleeding risk of 2.4%. For open anterior resection, the risk of symptomatic VTE was between 1.0% and 4.0% (3).

Colectomy involves the removal of part or all of the colon and can be performed for CRC. The decision to perform segmental or extended colectomy in CRC patients must consider the risk of metastasis and functional consequences of the surgery, age, and patients' wishes (16). The procedure can be done through open surgery or minimally invasive techniques such as laparoscopic or robotic-assisted surgery. The risk of symptomatic VTE is between 1.8% and 3.4%. The risk of VTE also varies by extent of bowel resection: minimally invasive and open left (1.4%) and right (1.9%) hemicolectomies have lower VTE risk compared with total proctocolectomies or total colectomies (laparoscopic 5.0% and open 5.4%) (3). As for bleeding requiring reintervention, minimally invasive right colectomy has the highest risk (1.5%), followed by minimally invasive colectomy for malignant disease (1.3%) (3).

Although patients with IBD are usually treated medically, surgery is required in patients who develop severe complications and those who are refractory to medical therapy (17). Subtotal colectomy with or without ileal pouch-anal anastomosis (IPAA) can be performed on IBD patients (18). The risk of symptomatic VTE is slightly higher than that of malignant disease with rates from 2.1% to 4.1% (3). Postoperative bleeding from the remnant rectum is a procedure-specific complication after subtotal colectomy (19). The reported incidence of postoperative bleeding from the remnant rectum ranges from 1.8% to 8.2% (19). Total proctocolectomy (TPC) is the surgical resection of the entire colon and rectum with or without perineal dissections (20). This procedure is associated with a higher risk of symptomatic VTE due to its extensive nature and the potential for postoperative complications (21). The reported VTE incidence ranges from 4.3% to 12.6%, while clinical relevant bleeding rates can reach up to 2.4%. The complexity of the surgery and the need for careful anastomosis contribute to these risks.

Rectopexy is a surgical procedure used to treat rectal prolapse by securing the rectum to the sacral promontory (22). This procedure can be performed through an abdominal or perineal approach. Abdominal surgery for rectal prolapse requires one larger incision or multiple smaller incisions (23). The risk of VTE in rectopexy is relatively lower compared to more extensive colorectal procedures (24). It has been reported that within 30 days post-operation, the incidence of symptomatic VTE for laparoscopic rectopexy is 0.4%, for open rectopexy is 0.6%, and for perineal rectopexy is 1.2%. Additionally, bleeding risks are also lower, with reintervention rates for bleeding around 0.4%. However, the anatomical challenges and the need for precise fixation can still contribute to potential complications.

Composite VTE includes both symptomatic and asymptomatic VTE, providing a comprehensive measure in clinical trials (25). A recent multicenter cohort study involving solid cancer patients reported a composite VTE rate of 11.4% at 6 mo during anticoagulant therapy (26). Similarly, a prospective cohort study (the CRC-VTE study) in China observed a high composite VTE rate of 11.2% in CRC patients following surgery (27). The rate of symptomatic VTE in this study was 2.5%, consistent with the rates in randomized controlled trials (3). Two decades ago, the incidence of composite VTE events in China was reported to be higher (up to 38%) compared with now (28), primarily because of discontinuation of anticoagulant therapy due to bleeding concerns and surgeons' unawareness of VTE prophylactic guidelines and adherence to consensus treatment (29). Attention must be paid to timing of diagnosing asymptomatic VTE because regularly scheduled clinical and radiographic examinations are difficult to perform in post-discharge course in real-world settings.

The variation in VTE and bleeding risks across colorectal procedures underscores the need for a tailored approach. For surgeries with a high risk of VTE, a more aggressive prophylaxis regimen and closer post-discharge surveillance may be warranted. For low-risk procedures, standard prophylaxis may be sufficient. Surgeons must integrate this procedure-specific risk profile into their perioperative anticoagulation planning.

Current guidelines recommend a PK-based approach for preoperative interruption of DOAC in patients undergoing elective colorectal surgery. For high bleeding risk surgery, preoperatice DOAC interruption should be maintained for four to five half-lives (35). Elimination half-lives of FXa inhibitors are 8–12 h in patients with creatinine clearance (CrCl) above 30 mL/min (36). Dabigatran has a higher dependence on kidney clearance, so its elimination half-life is 10–14 h in patients with a CrCl at above 50 mL/min and 18–24 h in patients with a CrCl of 30–49.9 mL/min (36). Specifically, the preoperative interruption interval should correspond to 8–96 h depending on different types of DOACs and hepatic/renal functions of individuals to ensure minimal or no residual anticoagulant effect at the time of surgery (35, 36).

The aforementioned strategy was investigated in the PAUSE study, which is a cohort study of 3,007 patients with AF and a CrCl >30 mL/min who were undergoing elective surgical or nonsurgical procedures while on DOAC therapy (37). The 30-d postoperative incidence rates of symptomatic VTE events and major bleeding were as follows: 0.48% [95% confidence interval (CI): 0.16%–1.40%] and 1.35% (95% CI: 0.0%–2.0%) in the apixaban cohort; 0.30% (95% CI: 0.06%–1.68%) and 0.9% (95% CI: 0.0%–1.73%) in the dabigatran cohort; and 0.09% (95% CI: 0.02%–0.87%) and 1.85% (95% CI: 0.0%–2.65%) in the rivaroxaban cohort (37). A retrospective single-center study evaluated the perioperative management of 525 patients on DOAC therapy undergoing elective surgery or procedures. Unlike the PAUSE study, perioperative DOAC management in this study was not standardized and was left to the discretion of the attending physicians. Using this approach, 2.4% of patients experienced major bleeding, and 0.8% had thromboembolic events, which were higher incidence rates compared to those observed in the PAUSE study (38).

Compared to warfarin, the anticoagulant effect of DOACs decreases more rapidly after interruption and takes effect more quickly upon resumption of administration (39), thereby eliminating the need for low molecular weight heparin (LMWH) bridging during the perioperative DOAC interruption period. Studies have demonstrated that the perioperative management of patients undergoing noncardiac surgery while using DOACs without heparin bridging is safe and feasible (36). In a prospective registry of 901 patients undergoing elective surgery while on DOAC therapy, perioperative bridging with LMWH was associated with an increased risk of major bleeding (OR = 4.6; 95% CI: 1.6–13.2), while it had no significant impact on thromboembolic outcomes (OR = 1.9; 95% CI: 0.7–5.4) (40). In a meta-analysis, compared to the nonbridging group, perioperative bridging with LMWH was associated with a threefold higher incidence of major bleeding (4.8%; 95% CI: 3.4%–6.2% vs. 1.6%; 95% CI: 1.2%–2.0%), with no difference in the pooled incidence of stroke/systemic embolism (0.4%; 95% CI: 0.1%–0.9% vs. 0.3%; 95%CI: 0.1%–0.4%) (41). Notably, Fujikawa et al. observed that, in patients undergoing elective gastrointestinal surgery, the incidence of postoperative major bleeding was significantly higher in those treated with DOACs along with heparin bridging compared to those treated with warfarin or DOACs alone (14.7% vs. 4.8% vs. 1.4%, P = 0.011) (42). Nevertheless, in selected high thrombotic risk patients (e.g., recent VTE within 3 months, mechanical heart valves, or prior thromboembolism during anticoagulation interruption), individualized consideration of bridging therapy may be warranted after multidisciplinary evaluation.

The timing of postoperative resumption of DOACs is based on the assessment of bleeding risk associated with surgery and hemostasis at the surgical site, including blood loss through surgical dressings and drains. Colorectal surgery is considered high-risk for bleeding, and guidelines recommend restarting DOACs 2 d after surgery (43). Flexibility in the timing of postoperative DOAC resumption is necessary, as the anticoagulant effect peaks 2–3 h after administration, which may increase the risk of postoperative bleeding, particularly relevant if patients experience greater than expected postoperative blood loss (36). If postoperative bleeding occurs or hemostasis at the surgical site is uncertain, resumption should be delayed (36).

In the PAUSE study, DOAC resumption was not initiated earlier than 48–72 h after surgery, with a high bleeding risk, which was associated with a 30-d postoperative major bleeding rate of 2.49% (95% CI: 0%–4.25%) in this population (37). Even in a subgroup of patients undergoing radical prostatectomy (identified as procedure with high thromboembolism risk and bleeding risk), the strategy of PAUSE achieved total rates of thromboembolic/bleeding complications within 30 d postoperatively as low as 3.7% (44).

Despite these advantages, anticoagulants remain the leading cause of emergency department visits due to adverse drug events, accounting for ∼14.90% (95% CI: 10.70%–19.10%) of all ADE-related visits in the USA from 2017 to 2019 (45). Clinical research over the past decade has revealed significant interpatient variability in DOAC plasma levels, raising concerns about the appropriateness of a “one size fits all” dosing strategy for DOACs (46, 47).

Quantitative measures for the direct assessment of DOACs' effects involve anti-FXa activity (for FXa inhibitors), dilute thrombin time (for dabigatran), ecarin thrombin time (for dabigatran), and drug plasma levels (48, 49). For the anti-FXa assay, FXa is added to plasma containing a FXa substrate (e.g., heparin) that is tagged with a chromophore. When the chromophore is cleaved by FXa, a color change results that is directly proportional to the levels of FXa present in the assay (50). The chromogenic anti-FXa assay for the anticoagulant effects of FXa inhibitors is precise, sensitive, and accurate with a plasma levels-dependent inhibition of FXa activity (correlation coefficient of 0.9669) (50). No effect is expected on the anti-FXa assay from dabigatran based on the mechanism of direct thrombin inhibition. Dabigatran prolongs the dTT in a linear dose-relationship and can accurately predict anticoagulation intensity (51). The dTT demonstrated a high correlation with dabigatran plasma levels when used with both in-house and Hemoclot-derived dabigatran calibrators (correlation coefficients of 0.9981 and 0.9982, respectively) (52). The ECT directly measures thrombin generation, with a correlation coefficient of 0.92 with dabigatran plasma levels (53). No effect is expected on dTT and ECT from this drug class based on the mechanism of direct FXa inhibition.

Few data exist regarding the effect of measuring preoperative DOAC levels on perioperative bleeding risk. Some have advocated that a preoperative DOAC level less than 50 ng/mL is a safe threshold to allow surgery to proceed, as it considered a minimal, clinically insignificant anticoagulant effect. And a DOAC level less than 30 ng/mL may be considered an undetectable level (54). For colorectal surgery, patients with multiple factors (e.g., age ≥75 years or renal insufficiency) interfering with PK of DOACs, or patients with uncertain time windows, may benefit from low residual DOAC levels, particularly for procedures associated with a high risk of bleeding (e.g., total proctocolectomy), where minimal or no anticoagulant effect is desired during surgery (37). Some studies have assessed preprocedural DOAC levels and clinical factors associated with higher residual levels (55, 56). A secondary analysis of PAUSE study was performed to identify risk factors associated with unsatisfied residual DOAC levels (55). In low-risk procedures, age ≥75 years, female sex, and CrCl >50 mL/min were associated with both DOAC levels ≥30 ng/mL and ≥50 ng/mL. Additionally, a DOAC interruption of 36 h was linked to levels ≥30 ng/mL, while standard DOAC dosing was associated with levels ≥50 ng/mL. For high-risk procedures, weight <70 kg, CrCl >50 mL/min, and standard DOAC dosing were associated with residual DOAC levels ≥30 ng/mL, whereas female sex was linked to levels ≥50 ng/mL (55). The CORIDA study, a prospective observational study involving 422 DOAC-treated patients undergoing elective procedures, measured preprocedural DOAC levels. Unlike the PAUSE study, it did not standardize perioperative DOAC management, resulting in a wide variation in DOAC interruption duration (1–218 h). The study identified the duration of DOAC interruption, CrCl <50 mL/min, and antiarrhythmic drug use as predictors of preprocedural DOAC levels ≥30 ng/mL, while patient age, sex, and weight were not significant predictors (56).

Meanwhile, the feasibility and cost-effectiveness of DOAC level monitoring remain important practical considerations. While the anti-FXa assay is becoming more accessible, its widespread implementation is constrained by factors including specialized laboratory infrastructure, trained personnel, and the limited availability of DOAC-specific calibrators. A survey of 46 specialized coagulation assays revealed that only 39% offered anti-Xa assays for rivaroxaban and 22% offered this assay for apixaban (57). Current guidelines applicable to China generally recommend a PK-based management strategy without mandatory preprocedural DOAC level measurement (58). This approach prioritizes a standardized interruption interval based on the drug's half-life, renal function, and procedural bleeding risk, reserving laboratory assessment for complex or high-risk scenarios (e.g., emergency surgery, severe renal impairment, or suspected overdose). Within this framework, greater emphasis is placed on the safety thresholds and interruption timelines. For elective surgery, a preprocedural DOAC level below 50 ng/mL is widely considered to represent a minimal anticoagulant effect associated with bleeding risk, while a level below 30 ng/mL is often regarded as negligible (54). To achieve this, a preoperative interruption period corresponding to four to five drug half-lives is recommended for procedures with high bleeding risk.

Patients treated with DOACs who require emergency surgery have a high risk of bleeding (17%–23%) and VTE (7%–16%) (59). Management decisions for patients taking DOACs who need emergency colorectal surgery involve multiple patient- and procedure-related factors, making it difficult to develop standardized management protocols (60, 61). The anticoagulant effect of a DOAC can be neutralized with DOAC specific reversal agents, including andexanet-α for apixaban, edoxaban and rivaroxaban or idarucizumab, a monoclonal antibody fragment that act as a specific reversal agent for dabigatran (62). Prothrombin complex concentrate (PCC) and activated PCC, which are nonspecific prohemostatic agents, can be used to reverse the effect of all DOACs (62). If DOAC level testing is available, a DOAC level at or above 50 ng/mL may necessitate the use of a DOAC reversal agent, whereas a level less than 50 ng/mL may allow the operation to proceed without a reversal intervention (63). If DOAC level testing is unavailable, DOAC reversal agent should be considered if the most recent DOAC dose was taken less than 48 h before the procedure (60).

Inflammatory bowel disease (IBD) is a chronic inflammatory disorder associated with a markedly prothrombotic state (64). The underlying inflammation elevates the risk of thromboembolism by 2–3 fold (65). This risk further amplified by active IBD disease, flares-up, surgery, steroid treatment, and hospitalization (66). Meanwhile, IBD patients may suffer gastrointestinal bleeding, especially those with ulcerative colitis (67). Anticoagulant therapy may increase the bleeding risk in IBD patients, which may consequently influence treatment decisions (68). For IBD patients on DOACs, a standardized PK-based interruption strategy is recommended (69). However, in patients with active severe colitis, recent significant gastrointestinal bleeding, or undergoing surgery involving extensive mucosal resection, a more conservative approach may be considered in consultation with a gastroenterologist and hematologist (69). Postoperative DOAC resumption should generally follow the high bleeding risk protocol (e.g., 48–72 h post-surgery after hemostasis is secured) (69). Bridging therapy is not routinely recommended due to its associated bleeding risk (70).

Frailty constitutes a high-risk category for perioperative DOAC management due to its amplification of both thrombotic and hemorrhagic complications (71). Frail patients present a complex challenge characterized by multimorbidity, polypharmacy, and reduced physiological reserve (72). Of particular concern is the heightened susceptibility to bleeding due to factors such as increased fall risk, acute kidney injury, and prevalent drugs interactions (73). Prolonged immobility and delayed recovery may extend the window of vulnerability to VTE (73). Therefore, a standard or more conservative DOAC interruption strategy is warranted. Preoperative interruption should follow PK timelines for high bleeding risk surgery (74). Postoperative resumption should be carefully timed, balancing thrombotic therapy needs against the heightened risk of spontaneous or traumatic hemorrhage (75). Factors such as dynamic renal function, nutritional status, and cognitive function must be serially evaluated to guide safe restarting of therapy.

Renal function is a critical determinant of DOACs clearance, particularly for dabigatran (∼85% renal excretion), and to a significant extent for edoxaban, rivaroxaban, and apixaban (76). Impaired renal excretion prolongs the half-life and elevates plasma levels, thereby increasing the risk of major bleeding during the perioperative period (76). For patients with moderate to severe renal impairment (e.g., CrCl 15–50 mL/min), a prolonged preoperative DOAC interruption period is mandatory. The duration must be tailored to the specific agent and the degree of renal dysfunction, often extending 24–48 h beyond the standard window recommended for normal renal function (77). In urgent situations or for patients with fluctuating renal function, laboratory monitoring can objectively assess residual anticoagulant activity prior to surgery (46, 63). Postoperative resumption must be highly individualized. Dose adjustment according to the guidelines based on current renal function is imperative upon restarting (76).