Half-Dose Rivaroxaban Cuts Bleeding After Surgery Without Raising Clot Risk, Study Finds

A single-center study from China suggests that patients who develop a small, isolated blood clot in the deep veins of the calf after urological surgery may be able to receive a much lower dose of the anticoagulant rivaroxaban than is conventionally prescribed, with substantially less bleeding and no detectable increase in clot progression. The research, published in Frontiers in Cardiovascular Medicine, compared a de-escalated regimen of 10 milligrams of rivaroxaban once daily against the standard maintenance dose of 20 milligrams once daily in patients diagnosed with isolated distal deep vein thrombosis, or IDDVT, after operations ranging from robotic prostatectomies to bladder tumor resections. The findings are provocative because they challenge a long-standing assumption that full-dose anticoagulation is the safest default in the fragile days and weeks following surgery, when the risks of hemorrhage and thrombosis are locked in a delicate tug-of-war.

Isolated distal deep vein thrombosis is a peculiar clinical entity. Unlike proximal deep vein thrombosis, which forms in the larger veins of the thigh and pelvis and carries a well-recognized risk of life-threatening pulmonary embolism, IDDVT is confined to the infrapopliteal veins below the knee. It has historically been regarded as relatively benign, yet modern registry data show a heterogeneous course: some clots resolve quietly, while others extend proximally, recur, or seed the lungs, particularly when left untreated. Because IDDVT is more often triggered by transient insults such as surgery or trauma rather than chronic systemic hypercoagulability, the evidence base built on non-surgical populations may simply not transfer to the postoperative ward. That gap is precisely where the new study positions itself.

The clinical dilemma is acute. Current guidelines generally recommend at least three months of anticoagulation for high-risk IDDVT, favoring direct oral anticoagulants such as rivaroxaban. But those recommendations rest largely on randomized trials that excluded patients at high risk of postoperative bleeding. Surgeons and physicians facing a fresh surgical wound in a hypervascular abdominopelvic field often hesitate to deploy full-dose anticoagulation, and real-world practice frequently drifts toward more conservative dosing. Whether that drift protects patients or leaves them under-protected against clot propagation has been unknown. The Dalian team, led by Hansheng Bai and senior authors Guangjun Fan and Si Wu, set out to quantify that trade-off using more than a decade of real-world data.

The researchers screened 302 patients who developed IDDVT after urological surgery at the Second Hospital of Dalian Medical University between January 2013 and April 2025. After exclusions for incomplete follow-up, pre-existing proximal clots, significant bleeding, or other indications for anticoagulation, 247 patients formed the initial cohort. Because dose selection in routine practice was made by treating physicians rather than by protocol, the investigators used propensity score matching to construct balanced comparison groups, pairing each patient on the 10-milligram regimen with a statistically similar patient on the 20-milligram regimen. The matching model incorporated an unusually thoughtful variable: the surgical approach itself, dichotomized as endoscopic procedures performed through natural orifices versus major structural surgeries involving abdominal or retroperitoneal dissection, as a surrogate for surgical trauma and bleeding risk. Year of surgery was also included to absorb temporal shifts in practice across the twelve-year window.

After matching, 152 patients remained, 76 in each arm. The groups were well balanced on age, body mass index, platelet count, D-dimer levels, comorbidities such as hypertension and diabetes, malignancy history, and the distribution of surgical approaches. The median time from surgery to clot diagnosis was one to two days, and anticoagulation typically began about three days later. Most patients completed three months of therapy, with a minority extended to six months when ultrasound failed to show resolution. The stage was thus set for a head-to-head comparison of two intensities of the same drug in a genuinely perioperative population.

The primary outcome was thrombus progression to proximal deep vein thrombosis or pulmonary embolism, analyzed with Fine–Gray competing-risk models that treated bleeding-related treatment discontinuation as a competing event. The cumulative incidence of progression at the end of follow-up was 11.84 percent in the low-dose group and 9.21 percent in the standard-dose group, a difference that was not statistically significant. The subdistribution hazard ratio of 1.29 came with a 95 percent confidence interval stretching from 0.48 to 3.46, and a P value of 0.613. Thrombus resolution told a similar story: roughly two-thirds of patients in each group achieved complete disappearance of the clot on follow-up ultrasound, with no significant difference between regimens.

Bleeding was where the regimens decisively parted ways. Clinically relevant non-major bleeding, classified according to International Society on Thrombosis and Haemostasis criteria, occurred in 6.58 percent of the low-dose group versus 19.74 percent of the standard-dose group, yielding a hazard ratio of 0.31 with a confidence interval of 0.11 to 0.85 and a P value of 0.023. The pattern was echoed in early hematuria: within the first month of treatment, dipstick-positive blood in the urine appeared in 23 patients on the standard dose compared with 12 on the low dose, a statistically significant difference. Bleeding also disrupted care far more often under full dosing, interrupting treatment in 19.74 percent of standard-dose patients versus 6.58 percent of low-dose patients. In a population where even macroscopic hematuria can force a dose reduction or outright discontinuation, that interruption burden is not a trivial footnote; it strikes directly at treatment continuity, the very mechanism by which anticoagulation fails to protect.

The authors are careful, almost emphatic, about what these results do not show. A wide confidence interval around the progression estimate, a modest sample size, nonrandomized treatment allocation, and the ever-present possibility of residual confounding mean the study cannot claim the two doses are equivalent in antithrombotic efficacy. Physicians may have preferentially prescribed lower doses to frailer patients, a classic confounding-by-indication problem that propensity matching can soften but never eliminate. Patients with significant renal impairment were excluded, limiting generalizability to older surgical patients, and follow-up ultrasound timing was not standardized, opening the door to surveillance bias. The findings are framed as hypothesis-generating rather than practice-changing, and the authors explicitly call for randomized non-inferiority trials to settle the question.

Nevertheless, the study lands in a converging evidence landscape. The EINSTEIN-CHOICE trial demonstrated that low-dose rivaroxaban can sustain antithrombotic protection during extended treatment in the broader venous thromboembolism population, and smaller studies in critically ill patients and in post-thoracic surgery IDDVT have reported comparable thrombotic outcomes with reduced dosing. A separate study of different rivaroxaban doses in IDDVT observed a significantly elevated bleeding risk with standard dosing, mirroring the Dalian results. Registry analyses add nuance: the RIETE registry has shown that concomitant superficial vein thrombosis raises pulmonary embolism risk, that cancer-associated IDDVT carries higher recurrence and bleeding risks, and, intriguingly, that low LDL cholesterol predicts anticoagulation-related bleeding, hinting that laboratory markers could one day sharpen individualized dosing decisions.

The pathophysiological logic is also coherent. Distal clots carry a smaller thrombus burden and lower rates of proximal extension than proximal thrombi, and when the dominant trigger is a transient surgical insult rather than systemic hypercoagulability, local factors may drive thrombosis more than circulating coagulation drive. In that setting, a full maintenance dose may deliver more bleeding than benefit during the early postoperative window when wounds are still healing. If randomized trials confirm the pattern seen here, the implications would ripple across surgical specialties: a cheaper, better-tolerated regimen that keeps patients on therapy instead of interrupting it, and a template for dose de-escalation in other high-bleeding-risk perioperative scenarios. For now, the message is one of cautious possibility, a signal that in the contest between clot and bleed after surgery, less anticoagulation may sometimes be exactly enough.

Subject of Research: Low-dose versus standard-dose rivaroxaban for postoperative isolated distal deep vein thrombosis

Article Title: De-escalated rivaroxaban in postoperative isolated distal deep vein thrombosis: a propensity score–matched study

Article References: Bai, H., Tang, Q., Jin, L., Yang, M., Fan, G., & Wu, S. (2026). De-escalated rivaroxaban in postoperative isolated distal deep vein thrombosis: a propensity score–matched study. Frontiers in Cardiovascular Medicine, 13, 1855986. https://doi.org/10.3389/fcvm.2026.1855986

Image Credits: AI Generated

DOI: 10.3389/fcvm.2026.1855986

Keywords: rivaroxaban, anticoagulation, deep vein thrombosis, isolated distal DVT, postoperative complications, bleeding risk, direct oral anticoagulants, propensity score matching, urological surgery, venous thromboembolism, dose de-escalation, thrombosis