10  Turning the Tide: Key Cases and Conundrums in Thrombosis Medicine

10.1 Session Overview

Session Details
Session Turning the Tide: Key Cases and Conundrums in Thrombosis Medicine
Speakers Lai Heng Lee, MBBS, MD; Ponlapat Rojnuckarin, MD, PhD; Jeffrey Weitz, MD
Affiliations Singapore General Hospital; Chulalongkorn University, Bangkok; McMaster University, Ontario
Time Day 2, 11:00 a.m.–12:15 p.m.

This case-based session is anchored by three longitudinal patient vignettes that together trace the full life cycle of cancer-associated thrombosis (CAT): risk factor assessment, agent selection, perioperative interruption, management of bleeds and thrombocytopenia, extended anticoagulation, drug–drug interactions, mechanical thrombectomy, and end-of-life care [slide p.1, p.3]. The learning objectives frame the discussion around (1) identifying tumor-, treatment-, and patient-related contributors to cancer VTE; (2) applying current diagnostic and risk-assessment tools; and (3) selecting evidence-based prevention and treatment plans that balance bleeding and thrombotic risk [slide p.3].

10.2 Speaker Spotlight

Lai Heng Lee, MBBS, MD heads the Department of Haematology at Singapore General Hospital and is a recognised authority on VTE epidemiology and anticoagulation practice in Asian populations [slide p.1]. Her talk walks through a 76-year-old Chinese man with overlapping haematologic and solid-tumour malignancies, using his trajectory to illustrate risk stratification, agent choice, perioperative management, and bleed response.

Ponlapat Rojnuckarin, MD, PhD is Professor of Medicine at Chulalongkorn University in Bangkok and a regional leader in coagulation disorders across Southeast Asia [slide p.1]. He takes on the hardest population in CAT—patients with primary brain tumours—using a 74-year-old man with high-grade glioma to demonstrate how to balance catastrophic bleeding risk against unavoidable thrombotic risk.

Jeffrey Weitz, MD is Professor of Medicine at McMaster University and a pioneer in novel anticoagulant development [slide p.1]. He presents a 71-year-old man with metastatic castration-sensitive prostate cancer on androgen-receptor-axis therapy, using him to work through DOAC–ARAT drug interactions, mechanical thrombectomy evidence, extended anticoagulation, concurrent antiplatelet exposure, and transitioning to palliative care.

10.3 Case 1 — Mr TYC: Cancer VTE, Procedures, and Bleeds on Treatment

10.3.1 Patient background

Mr TYC is a 76-year-old Chinese man with hypertension, obesity, and right-knee osteoarthritis [slide p.4]. His oncologic history spans two primary malignancies: atypical chronic lymphocytic leukaemia and splenic marginal zone lymphoma treated with six cycles of rituximab–bendamustine in 2019, and a stage 2a rectosigmoid adenocarcinoma resected in 2022 with adjuvant chemotherapy [slide p.4]. In April 2024 he developed lymphoproliferative progression with rising LDH, enlarging nodes, splenomegaly, and new FDG avidity, managed with watch-and-wait given a low disease burden [slide p.4].

In July 2025 he was admitted for small-bowel obstruction from adhesions, underwent adhesiolysis with subcutaneous enoxaparin prophylaxis, then developed hospital-acquired pneumonia and an eGFR drop from 73 to 43; restaging CT showed new pelvic nodules invading the rectal wall and bladder with fresh pelvic and mesorectal lymphadenopathy [slide p.5]. Two weeks after discharge, in August 2025, he re-presented with right lower-limb swelling and pain extending to the proximal thigh. Doppler ultrasound confirmed acute right common iliac vein thrombosis, and MRI showed compression of the right iliac vein by surrounding pelvic adenopathy—a mechanical contribution on top of the biological VTE risk [slide p.6].

10.3.2 Risk factors for DVT in haematologic malignancy

Dr. Lee anchored the discussion in the ASH 2025 education session Anticoagulants in hematologic malignancies: what is the data? [slide p.8]. VTE incidence in hematologic malignancies ranges from <5% in chronic leukaemias to 59.5% in primary CNS lymphoma, with 8–12% in aggressive lymphoma and 10–26% in multiple myeloma [slide p.8]. Malignancy-related drivers include aberrant tissue factor expression, inflammatory cytokine–driven FVIII/VWF elevation, endothelial injury, DIC, platelet activation, impaired fibrinolysis, and increased plasma viscosity [slide p.8]. Patient-related factors overlap with Virchow’s triad—older age, hormonal therapy in females, prior VTE, ECOG ≥2, immobility, central venous catheters, hypertension/renal/lung/diabetes comorbidity, obesity, and systemic infection [slide p.9]. Mr TYC’s additional risks were the presence of a second active cancer and mechanical iliac vein compression by pelvic masses [slide p.9].

VTE risk scores used in haematologic malignancies [slide p.10]
Risk score Cancer setting Key components Predictive performance Limitations
Khorana (2008) Solid tumours and lymphoma Cancer site, platelets ≥350, Hgb <100, ESA, WBC >11, BMI ≥35 6-month VTE ~6.7% at score ≥3; C-stat 0.7 Moderate discrimination; poor in lymphoma; not leukaemia-specific
Vienna V-CATS Mixed (mostly solid) Khorana factors + D-dimer + soluble P-selectin Calculated VTE probability; 87.5–100% with modified V-CATS Complex; proof-of-concept only in heme malignancy
THROLY (2016) Lymphoma Prior thrombosis, BMI ≥30, mediastinal involvement, reduced mobility, extranodal involvement, ANC <1, Hgb <100 C-stat ~0.55 Limited external validation; not universally adopted
Lymphoma VTE Nomogram (Liang 2023) Lymphoma Age, male sex, platelets, D-dimer, number of chemotherapy cycles C-stat ~0.78 Chinese cohort; limited external validation

10.3.3 Choice of anticoagulant

For cancer-associated thrombosis, FXa inhibitors—apixaban and rivaroxaban—have displaced LMWH as standard of care on the strength of CARAVAGGIO and SELECT-D, and ASH now recommends an oral FXa inhibitor over LMWH for initial treatment [slide p.12]. The nuances matter:

  • Rivaroxaban (SELECT-D): similar major bleeding to LMWH overall, but higher clinically relevant non-major bleeding, concentrated in upper GI cancers [slide p.12].
  • Apixaban (CARAVAGGIO post hoc): no difference in bleeding in GI cancers versus LMWH [slide p.12]. However, the prospective observational study by Houghton et al. showed increased major bleeding with apixaban vs enoxaparin in luminal GI cancers, with lower CRNMB versus rivaroxaban [slide p.12].
  • LMWH retains its role in unresected luminal upper GI cancers complicated by VTE [slide p.12].

10.3.4 VICTORIE: real-world DOAC vs LMWH in cancer VTE

The VICTORIE study (Oral Communication #332, ASH 2025) is a multi-country retrospective cohort pooling real-world registry data from Norway, Sweden, Finland, the UK, and Germany to compare safety and effectiveness of DOACs versus LMWH in cancer-associated VTE [slide p.13]. Adults with incident VTE, active cancer (diagnosed in the 6 months prior), and a first anticoagulant dispensation within 30 days of the index event were included; prior VTE, prior anticoagulation, atrial fibrillation/valve disease, IVC filter insertion, and pregnancy were excluded [slide p.15]. Propensity score methods with inverse probability of treatment weighting (IPTW) were applied within each country-specific dataset to control confounding [slide p.17, p.18].

After IPTW, the apixaban–LMWH weighted comparison comprised 30,002 vs 2,892 patients and the rivaroxaban–LMWH weighted comparison 29,976 vs 4,321 patients [slide p.18]. Baseline characteristics were well balanced; mean age 67–68 years, mean Charlson Comorbidity Index ~5, with approximately one-third of patients having GI cancer [slide p.19, p.20].

VICTORIE bleeding hazard ratios vs LMWH [slide p.21]
Bleed outcome Apixaban vs LMWH HR (95% CI) Rivaroxaban vs LMWH HR (95% CI)
Overall bleed 0.67 (0.53–0.86) 0.98 (0.83–1.15)
GI bleed 0.89 (0.61–1.29) 0.86 (0.47–1.58)
Intracranial haemorrhage 0.67 (0.31–1.44) 0.32 (0.13–0.82)
Other bleed 0.64 (0.41–0.998) 0.89 (0.74–1.08)

Recurrent VTE was underpowered for comparative analysis. Apixaban was associated with lower overall and “other” bleeding than LMWH, with similar GI and ICH rates. Rivaroxaban showed similar overall, GI, and other bleeding to LMWH, but a significantly lower risk of intracranial haemorrhage [slide p.21, p.22]. These findings support DOACs as a safe LMWH alternative in real-world cancer-VTE populations where trial evidence remains limited [slide p.22].

10.3.5 Suspending anticoagulation for procedures

Mr TYC was started on therapeutic enoxaparin 80 mg BD for his new iliac DVT and developed small-volume bright-red PR bleeding, attributed to external haemorrhoids, without a significant haemoglobin drop. Enoxaparin was continued with plans for elective IR-guided biopsy of a perianal mass [slide p.23]. This raised the fundamental dilemma of suspending anticoagulation in a patient at very high recurrence risk.

The ASH 2025 education session on anticoagulation in the patient who bleeds and clots simultaneously offered a structured algorithm [slide p.25]: when a patient with newly diagnosed acute VTE needs surgical intervention, discontinuing anticoagulation within 4–6 weeks of the index VTE carries a recurrent VTE risk >20% per year. If possible, delay surgery for at least 4–6 weeks of therapeutic anticoagulation. When delay is impossible, minimise the perioperative window, bridge with unfractionated heparin or LMWH to shorten that window, restart early with prophylactic-dose anticoagulation post-operatively, and consider a temporary IVC filter [slide p.25].

10.3.6 Bleed management, thrombocytopenia, and transition

Biopsy was performed two weeks later; enoxaparin was restarted at a reduced 40 mg BD dose. He developed worsening renal function (creatinine 90 → 160) from right ureteric obstruction by pelvic masses, requiring temporary enoxaparin suspension and right-sided percutaneous nephrostomy [slide p.26]. Histology confirmed diffuse large B-cell lymphoma, and R-CHOP was initiated [slide p.26]. Enoxaparin was restarted at 40 mg OD after PCN removal but gross haematuria developed with a falling haemoglobin, and with counts dropping on R-CHOP, enoxaparin was again suspended [slide p.27].

The anticoagulation–thrombocytopenia framework (Hematology 2025 education program) provides graded guidance [slide p.29, p.30]:

  • Primary VTE prophylaxis: LMWH or DOAC if platelets ≥50 × 10⁹/L; higher thresholds (≥80) in select high-risk settings for DOACs; avoid pharmacologic prophylaxis <50 [slide p.29].
  • Secondary prophylaxis / maintenance: full-dose AC if platelets >50; reduced-dose LMWH 25–50; hold if <25 [slide p.29].
  • Acute VTE treatment: full-dose LMWH or DOAC if platelets >50; half-dose LMWH (no DOAC/VKA) 25–50; hold or dose-reduce with platelet support if <25 [slide p.29].
  • Practical levers: transfuse to ≥30–50 to support AC (≥40–50 for PE or mechanical valve); hold if bleeding risk is high or on grade-4 TP; consider TPO agonists (eltrombopag, romiplostim, avatrombopag) for prolonged chemotherapy- or CAR-T-induced TP when AC is essential; resume full-dose AC when platelets recover to ≥40–50 [slide p.30].

After PCN removal, enoxaparin was restarted at 40 mg OD, slowly escalated to 40 mg BD, and then transitioned to apixaban 5 mg BID. Apixaban was held during severe nadir thrombocytopenia, and he completed six cycles of R-CHOP without further bleeding [slide p.31].

10.3.7 ADAPT-BLEED: dynamic ML-based bleeding risk scoring

The case closed with Dr. Lee highlighting ADAPT-BLEED (Likvornik et al., Abstract #abs25-10073), a dynamic, context-aware bleeding risk score for patients receiving anticoagulation [slide p.32]. The multicentre Israeli study used EHR data from six academic centres over 12 years (2012–2024), applying a 90-day landmark approach to eliminate immortal-time bias [slide p.33]. The cohort included 163,596 patients (median age 74, 76.7% on enoxaparin, 19.9% on DOACs, 18.1% on concurrent antiplatelet), with major bleeding occurring in 1.5% at 6 months, 2.5% at 12 months, and 5.0% at 36 months [slide p.34].

An iterative clinical–ML approach (EHR extraction → 450+ candidate feature engineering → XGBoost training → SHAP analysis → clinical validation) yielded a 19-feature practical point system [slide p.35, p.36]:

ADAPT-BLEED point system [slide p.36]
Feature category Examples Point range
Demographics Age, gender 0–3
Medical history Stroke, prior bleeding, cancer 0–5
Chronic conditions Hypertension, CKD, diabetes, anaemia 0–3
Laboratory values Haemoglobin, creatinine, INR 0–5
Anticoagulant Warfarin, rivaroxaban, apixaban, dabigatran, enoxaparin 0–2
Antiplatelet Aspirin, clopidogrel 0–1

ADAPT-BLEED achieved an AUC of 0.72 vs 0.61 for HAS-BLED, an 18% improvement, with 4,000 bleeding events in the 163,596-patient cohort [slide p.36]. A worked example illustrated the point: Mr Levi, an 82-year-old on warfarin with stroke history and diabetes, scored 12 at baseline (moderate risk, ~3.5% annual bleeding); three months later his score rose to 7/10 after haemoglobin dropped to 9.2 g/dL, creatinine to 1.9 mg/dL, and INR to 3.2, reclassifying him as very high risk (~7.0% annual bleeding) [slide p.37]. The key clinical implication is that static risk scores miss dynamic changes; context-aware scoring with real-time lab integration adjusts risk weights based on evolving patient profiles and can be embedded into clinical workflows via a web calculator [slide p.38, p.39].

Mr TYC — Key Teaching Points
  1. Cancer-associated VTE risk is multifactorial: haematologic malignancy biology, solid tumour burden, mechanical compression, hospitalisation, infection, and procedures all stack.
  2. DOACs (apixaban or rivaroxaban) are preferred for most CAT—the VICTORIE real-world data are reassuring on bleeding—but luminal upper-GI cancers remain an LMWH indication.
  3. Anticoagulation held <4–6 weeks after acute VTE carries >20%/year recurrent-VTE risk; minimise the perioperative window and consider bridging.
  4. Use the platelet-count-stratified thrombocytopenia framework and dose-reduce rather than stop when possible; TPO agonists are an emerging adjunct.
  5. Static bleeding scores (HAS-BLED) are outperformed by dynamic, EHR-integrated ML tools like ADAPT-BLEED (AUC 0.72 vs 0.61).

10.4 Case 2 — Mr M: Brain Tumour with the Hardest Bleeding–Clotting Balance

10.4.1 A 74-year-old with high-grade glioma

Mr M presented in June 2025 with three weeks of dysarthria; brain MRI showed a left-hemisphere tumour with small intra-tumoural bleeding [slide p.41]. He underwent craniotomy in July with pathology confirming high-grade glioma (CNS WHO grade 4, IDH-wildtype) and received concurrent radiotherapy with temozolomide. In September 2025 he was hospitalised with pneumonia and seizures, raising the admission-era question: should VTE prophylaxis be given? [slide p.41]

10.4.2 Why brain tumours sit in a class of their own

Primary brain tumours have a VTE rate 2–3 times higher than other cancers, with high-grade glioma carrying the highest incidence at ~16% [slide p.42]. High-grade glioma overexpresses podoplanin (PDPN), which directly activates platelets via CLEC-2—a unique procoagulant mechanism [slide p.54]. IDH1 mutations and CDKN2A deletions downmodulate PDPN and tissue factor, while circulating tumour cells are found in ~20% of gliomas [slide p.54]. The counterbalancing reality is that these patients also face the highest baseline intracranial haemorrhage risk: brain metastases from melanoma, renal cell carcinoma, thyroid cancer, and choriocarcinoma are especially prone to bleeding, and high-grade glioma has increased ICH risk after anticoagulation initiation [slide p.42]. Biopsy, surgery, and bevacizumab add further VTE and bleeding pressures [slide p.42].

10.4.3 A cancer-specific inpatient VTE risk assessment model

Dr. Rojnuckarin highlighted Martin et al.’s new hospital-acquired VTE (HA-VTE) risk assessment model for active cancer patients on medical services [slide p.43, p.44]. Developed from the MITH dataset (6 US hospital systems, 14 hospitals, 2016–2022), the study compared three approaches [slide p.45, p.46, p.47]:

  1. Model 1: The existing general-medical MITH RAM (history of VTE, creatinine >2 or dialysis, anaemia, high RDW, low sodium, malnutrition, anticoagulation level within 24 h of admission) [slide p.46].
  2. Model 2: MITH model refit with addition of a cancer-type variable [slide p.47].
  3. Model 3: A de novo RAM.

The validation cohort had 58,124 admissions; top variables in the de novo model included prior VTE (OR 2.95), malnutrition (OR 1.52), hematologic malignancy excluding aggressive lymphoma (OR 1.42), lung cancer (OR 1.39), inflammatory comorbidity (OR 1.52), and full anticoagulation at admission (OR 1.57); age ≥75 was counterintuitively protective (OR 0.57) [slide p.48, p.49]. Performance by AUC was 0.58 for MITH alone, 0.60 for MITH + cancer type, and 0.64 for the de novo cancer-specific RAM in validation [slide p.50]. The message: a cancer-specific inpatient RAM outperforms a general medical RAM, and the predictor set is genuinely different [slide p.51].

A second critical observation came from Li et al. (Blood 2024): bleeding in real-world cancer-VTE cohorts is roughly five-fold higher than in clinical-trial placebo arms (~5% vs ~1% major bleeding at 6 months; ~6% vs ~3% CRNMB), with ~10% case-fatality in real-world major bleeds [slide p.52]. Trial stringency under-represents the true at-risk population, and fine-tuned risk prediction is needed before widespread prophylaxis.

10.4.4 Isolated distal DVT in cancer is NOT benign

Mr M returned in December 2025 with aphasia and right hemiparesis, brain oedema, old intratumoural haemorrhage, depression-driven immobility, and a newly diagnosed right peroneal vein thrombosis—an isolated distal DVT [slide p.53]. The question: anticoagulate or not?

Data make the answer clearer than it used to be. A Righini et al. study comparing serial proximal ultrasound vs single complete US for distal DVT detection showed equivalent 3-month VTE rates (~1%), but this was in general populations—and ACCP recommends anticoagulation in high-risk distal DVT defined by active cancer, positive D-dimer, extensive thrombosis, proximity to proximal veins, absence of reversible provoking factors, VTE history, or inpatient status [slide p.55]. The RIETE registry data (>90% of whom received anticoagulants) demonstrates that cancer distal DVT is not benign:

RIETE registry outcomes for cancer vs non-cancer distal DVT [slide p.56]
Outcome (% with 95% CI) Cancer Proximal DVT (n=5,196) Cancer Distal DVT (n=886) Non-cancer Distal DVT (n=5,974)
3-month recurrent VTE 4.0 (3.5–4.6) 3.6 (2.6–5.1) 1.1 (0.9–1.4)
3-month major bleeding 3.6 (3.2–4.2) 2.7 (1.8–4.0) 0.7 (0.6–1.0)
3-month mortality 22.2 (21–23) 17.4 (15–20) 1.22 (1.0–1.5)
1-year recurrent VTE 6.5 (5.9–7.2) 6.3 (4.9–8.1) 2.4 (2.1–2.9)
1-year major bleeding 4.6 (4.0–5.2) 3.5 (2.5–4.9) 1.1 (0.8–1.4)
1-year mortality 35.9 (35–37) 30.7 (28–34) 1.8 (1.5–2.1)

Cancer distal DVT carries recurrence, bleeding, and mortality rates that are close to cancer proximal DVT and an order of magnitude higher than non-cancer distal DVT—fully justifying therapeutic anticoagulation [slide p.56].

10.4.5 The ICH algorithm for VTE in brain cancer

Leader, Wilcox, and Zwicker’s framework (Blood 2024;144:1781–90) guides decision-making when brain-tumour patients have recent tumour-related ICH and an incident VTE [slide p.57]:

  • PE or proximal lower-extremity DVT: place IVC filter.
  • Isolated distal DVT or isolated subsegmental PE: hold anticoagulation, serial lower-extremity ultrasound.
  • All other VTE sites: assess ICH volume. If >10 mL or multifocal (high ICH expansion risk), hold AC. If lower volume (intermediate risk) but very high thromboembolic risk, consider prophylactic-dose AC starting 2–4 days after diagnosis; non-contrast head CT at 24–48 hours; escalate to full dose if stable on serial imaging (typically 10–14 days).

Choices map to three scenarios: high bleeding risk → hold AC and follow; high thrombotic risk → full AC; both high → low-dose AC with follow-up escalation, or temporary IVC filter. Periodic reassessment is essential—risks are not static [slide p.57].

10.4.6 DOAC vs LMWH in brain cancer

Pooled data (Zwicker 2016, Profidia 2020, Iyengar 2024) show that in primary brain cancer, anticoagulation increases bleeding 3.7-fold versus no anticoagulation, but DOACs reduce bleeding to ~0.35× of LMWH—a striking advantage unique to primary brain tumours [slide p.58]. In metastatic brain cancer, neither anticoagulation vs none (1.07× NS) nor DOAC vs LMWH (1.05× NS) reaches significance. The practical recommendation: apixaban 5 mg BID for at least 3 months [slide p.58].

Mr M received therapeutic apixaban, his DVT resolved, and in March 2026 bevacizumab was started for disease progression—raising the next dilemma [slide p.59]. A meta-analysis of 22 RCTs (N=20,050) showed bevacizumab significantly increased arterial thrombosis (RR 1.37), venous thrombosis (RR 1.29), and major bleeding (RR 2.74, 95% CI 2.38–3.15) via endothelial injury, underlying cancer substrate, and impaired wound healing [slide p.60]. This anti-VEGF effect tips the bleeding–clotting balance further and demands vigilance.

10.4.7 API-CAT and extended reduced-dose DOACs

The practical question for Mr M beyond 6 months: continue full dose? Reduce? API-CAT (Mahé et al., NEJM 2025;392:1363–73) randomised 1,766 patients with active cancer and VTE who had completed ≥6 months of anticoagulant treatment to reduced-dose apixaban 2.5 mg BID vs full-dose 5 mg BID, stratified by cancer site and index VTE type, with a median 12-month treatment window [slide p.61, p.82]. The hierarchical design tested (1) non-inferiority of reduced dose for recurrent VTE prevention and (2) superiority of reduced dose for clinically relevant bleeding.

API-CAT key outcomes [slide p.83, p.84, p.85]
Outcome Reduced (2.5 mg BID) Full (5 mg BID) Subhazard / Hazard ratio (95% CI) p-value
Recurrent VTE 18 (2.1%) 24 (2.8%) 0.76 (0.41–1.41) 0.001 for non-inferiority
Clinically relevant bleeding (MB + CRNMB) 102 (12.1%) 136 (15.6%) 0.75 (0.58–0.97) 0.03 for superiority
Major bleeding 24 (2.9%) 37 (4.3%) 0.66 (0.40–1.10) –

A meta-analysis of three RCTs (N=2,361) combining apixaban 2.5 mg BID and rivaroxaban 10 mg/day confirmed the direction: reduced dose lowered the composite of VTE recurrence + major bleeding (RR 0.77, 95% CI 0.64–0.93) and reduced major + CRNMB (RR 0.76, 95% CI 0.62–0.93), with no difference in recurrence, major bleeding, or mortality alone [slide p.62]. Dr. Rojnuckarin chose reduced-dose apixaban for Mr M, rationalising that the tumour remained active while no clinically relevant bleeding had occurred—reflecting the emerging consensus that extended reduced-dose DOAC is effective and safe in the active-cancer, 6-month-plus setting [slide p.62, p.63].

Mr M — Key Teaching Points
  1. High-grade glioma has the highest VTE rate of any cancer (~16%) and the highest ICH risk after AC—podoplanin-driven platelet activation is the unique driver.
  2. Cancer-specific inpatient VTE RAMs outperform general-medical RAMs; real-world bleeding is ~5× trial placebo arms.
  3. Cancer distal DVT is not benign: RIETE 3-month mortality 17.4% and 1-year 30.7%—treat.
  4. DOACs (apixaban) halve bleeding vs LMWH in primary brain cancer—an under-appreciated advantage. At least 3 months of apixaban 5 mg BID.
  5. After 6 months of full-dose, reduced-dose apixaban 2.5 mg BID is non-inferior for recurrence and superior for bleeding (API-CAT, NEJM 2025).
  6. Bevacizumab increases bleeding 2.74-fold—recalibrate anticoagulation intensity when it is added.

10.5 Case 3 — Mr MT: Drug–Drug Interactions, Thrombectomy, and End-of-Life Care

10.5.1 Prostate cancer, apalutamide, and segmental PE

Mr MT is a 71-year-old with metastatic castration-sensitive prostate cancer on androgen-deprivation therapy (ADT) plus apalutamide, with no other comorbidities. Baseline: weight 65 kg, haemoglobin 110 g/L, platelets 135, creatinine 157 µmol/L, creatinine clearance 35 mL/min. He presents with a new diagnosis of multiple segmental PE [slide p.66].

The apparent menu of initial anticoagulation options [slide p.67]—apixaban 10 mg BID × 10 days then 5 mg BID; LMWH × 5 days then dabigatran 150 mg BID or edoxaban 30 mg daily; rivaroxaban 15 mg BID × 21 days then 20 mg daily; dalteparin 200 U/kg × 1 month then 175 U/kg; enoxaparin 1 mg/kg BID or 1.5 mg/kg daily; tinzaparin 175 U/kg daily—must be filtered through current guidelines and, critically, drug–drug interactions.

10.5.2 ASH 2021 guidelines and the DOAC/LMWH decision

ASH 2021 Recommendation 20 suggests DOAC (apixaban or rivaroxaban) or LMWH for initial CAT treatment (conditional, very low certainty), and Recommendation 23 suggests DOAC (apixaban, edoxaban, or rivaroxaban) over LMWH for 3–6 month treatment (conditional, low certainty) [slide p.68]. The fine print flags GI cancer as a DOAC caution and demands individualised assessment of DDI, bleeding, patient preference, and cost.

Carrier et al.’s algorithm layers concrete decision points [slide p.69]: if high bleeding risk (prior bleeding, high-risk GI lesion, platelets <50, renal/liver impairment, concurrent antiplatelets), unresected luminal GI or GU cancer, or significant DOAC drug–drug interactions (strong CYP3A4 or P-gp modulators) → LMWH preferred; otherwise DOAC preferred.

10.5.3 Apalutamide–DOAC: a catastrophic interaction

Apalutamide is a strong CYP3A4 and P-gp inducer. Leblanc et al. (Cancers 2024) systematically mapped DOAC–ARAT interactions; the apalutamide column is uniformly red [slide p.71]:

  • Apixaban + apalutamide: major/catastrophic → avoid this pair.
  • Rivaroxaban + apalutamide: major/catastrophic → avoid.
  • Edoxaban + apalutamide: moderate → avoid.
  • Dabigatran + apalutamide: major/catastrophic → avoid.
  • Warfarin + apalutamide: major/catastrophic → modify dose and monitor.

Darolutamide, in contrast, is a much cleaner partner (mostly negligible/minor). With Mr MT already committed to apalutamide, the DOACs were effectively off the table, and tinzaparin 12,000 IU SC daily was selected as therapeutic LMWH [slide p.77]. Three months later, when his cancer progressed and apalutamide was discontinued, LMWH was transitioned to apixaban 5 mg PO BID [slide p.77].

10.5.4 Would mechanical thrombectomy have helped?

If Mr MT had presented with massive PE or extensive iliofemoral DVT, would thrombectomy be indicated? Djulbegovic et al. (Abstract #332, ASH 2025) reported a consecutive MUSC series of 73 thrombectomy procedures (DVT 37%, PE 49%, other 14%) performed July 2023–December 2024 [slide p.73, p.74, p.75]. VTE recurrence occurred in 21/73 (29%) by categorical analysis, with 86% of recurrences within 90 days and 81% at the same site. Kaplan–Meier time-to-event analysis showed recurrence rising to ~45%, higher than the categorical estimate once censoring and follow-up length were accounted for. Six patients had two recurrences and one had six [slide p.75]. The message: mechanical thrombectomy has a substantial long-term VTE recurrence risk, concentrated early and at the same site.

The DEFIANCE trial (Abramowitz et al., Am Heart J 2025;281:92–102) is the first RCT of mechanical thrombectomy vs anticoagulation alone for unilateral iliofemoral DVT in patients with Villalta score >9, no pre-existing IVC filter or stent, and symptom onset ≤12 weeks; it will inform whether a thrombectomy-based interventional approach reduces post-thrombotic syndrome morbidity [slide p.76]. Until those data mature, caution around routine thrombectomy use is warranted.

10.5.5 How long to continue anticoagulation?

After 6 months of full-dose AC, the decision tree balances ongoing cancer activity against bleeding risk. Sanfilippo et al.’s framework [slide p.79] lays out factors arguing for continuation (ongoing active cancer, metastatic/progressive disease, ongoing systemic therapy, high-risk tumour types [pancreatic, gastric, lung], prior VTE, symptomatic PE as index event, residual vein thrombosis, recurrent VTE on AC) versus termination (cancer in complete remission, no ongoing therapy, no distant metastasis, low-risk tumour types [breast, prostate], high bleeding risk, life-threatening bleed, severe thrombocytopenia, end-of-life scenario). Periodic risk-benefit evaluation is the governing principle.

DALTECAN and TiCAT studies showed that with therapeutic LMWH, major bleeding concentrates in month 1 (DALTECAN 3.6%; TiCAT 0.8%) and falls thereafter; recurrent VTE similarly concentrates early [slide p.80]. For Mr MT, Dr. Weitz chose to transition to apixaban 2.5 mg PO BID for secondary prevention [slide p.96].

10.5.6 Therapeutic vs prophylactic dose apixaban (EVE trial)

The EVE trial (McBane et al., JTH 2024;22:1704–14), N=360, randomised cancer VTE patients to extended apixaban 2.5 mg BID vs 5 mg BID [slide p.81]:

  • Major + CRNMB: 8.9% (2.5 mg) vs 12.2% (5 mg); HR 0.72 (95% CI 0.38–1.37; p=0.39).
  • Major bleeding: 2.8% vs 2.2%; HR 1.26 (p=0.73).
  • Recurrent VTE + ATE: 5.0% vs 5.0%; HR 1.0 (p=1.00).

EVE suggested—but did not prove—non-inferior efficacy with numerically lower bleeding at the reduced dose, setting the stage for API-CAT.

10.5.7 Apixaban or rivaroxaban for extended AC? The RENOVE post-hoc

The RENOVE trial (Couturaud et al., Lancet 2025;405:725–35) enrolled 2,774 VTE patients at high recurrence risk after 6–24 months of AC and randomised them to reduced-dose (apixaban 2.5 BID or rivaroxaban 10 OD) vs full-dose DOAC, with a median 36-month follow-up [slide p.88]. The post-hoc analysis presented by Girard et al. (ASH 2025) compared apixaban (n=1,255) vs rivaroxaban (n=1,513) across crude, adjusted, and IPTW analyses [slide p.87, p.89, p.90].

RENOVE post-hoc: apixaban vs rivaroxaban in extended AC [slide p.91, p.93, p.94]
Outcome (5-year cumulative incidence) Rivaroxaban Apixaban Adjusted HR (95% CI)
Clinically relevant bleeding (all) 11.7% 13.7% 0.90 (0.70–1.17)
Major bleeding (all) 2.5% 3.9% 0.82 (0.46–1.44)
Recurrent VTE (all) 2.2% 1.7% 1.15 (0.57–2.31)

Crude, adjusted, and IPTW analyses all pointed to comparable risks of clinically relevant bleeding and recurrent VTE [slide p.95]. However, the negative-control outcome (serious adverse events unrelated to bleeding or VTE) showed a significantly lower rate with rivaroxaban, indicating residual confounding from treatment channelling. The conclusion: the two DOACs are pragmatically equivalent for extended VTE prevention, but non-randomised allocation likely influenced observed differences [slide p.95].

10.5.8 Concurrent antiplatelet therapy: the AVERT post-hoc

Mr MT later developed an NSTEMI (medically managed) and was started on aspirin 81 mg daily alongside apixaban 2.5 mg BID. Does this change his bleeding or recurrence risk? [slide p.96]

The AVERT post-hoc (Saad et al., ASH 2025) analysed 288 apixaban and 275 placebo patients in the original thromboprophylaxis trial, with 34.4% (apixaban arm) and 30.9% (placebo arm) on concurrent antiplatelet agents, NSAIDs, or both [slide p.98]. In the apixaban arm, concurrent antiplatelet agent use was associated with [slide p.100]:

  • Clinically relevant bleeding: HR 2.02 (95% CI 1.29–3.18) ↑
  • Major bleeding: HR 0.74 (0.16–3.45) NS
  • CRNMB: HR 2.63 (1.59–4.35) ↑
  • VTE: HR 0.73 (0.21–2.53) NS

The conclusion: combining drugs with antiplatelet properties with prophylactic-dose apixaban roughly doubles bleeding risk without any efficacy trade-off; clinicians should re-evaluate the indication for each antiplatelet/NSAID when initiating thromboprophylaxis in cancer patients [slide p.101]. For Mr MT, the ACS indication for aspirin was compelling, but the bleeding burden must be acknowledged and monitored.

10.5.9 End-of-life anticoagulation

Mr MT ultimately required rehospitalisation for general deterioration and was transferred to palliative care. Should AC be continued? [slide p.102, p.103]

Noble et al. (BMJ Supportive & Palliative Care 2019) followed 457 CAT patients; 214 died within 2 years [slide p.104]. CRNMB occurred in 5.5%, but bleeding events clustered near death: 11/12 (92%) occurred in the final 7 days. There was no evidence of recurrent VTE or major bleeding signal earlier in the dying trajectory. The options range from stopping AC entirely, switching to LMWH at prophylactic doses (dalteparin 5,000 IU; enoxaparin 40 mg; tinzaparin 4,500 IU), continuing DOAC at prophylactic doses (apixaban 2.5 mg BID; edoxaban 30 mg; rivaroxaban 10 mg), or maintaining therapeutic dosing [slide p.103]. The pragmatic choice is de-escalation—prophylactic-dose LMWH or DOAC—with discontinuation in the final days if bleeding or distress appears.

Mr MT — Key Teaching Points
  1. Apalutamide + DOAC is catastrophic—always screen ARATs against the DOAC interaction grid (Leblanc et al., Cancers 2024). Darolutamide is the cleaner ARAT partner.
  2. Mechanical thrombectomy carries substantial same-site VTE recurrence (~29% categorical, ~45% by KM); reserve for carefully selected patients. DEFIANCE RCT will clarify the PTS question.
  3. Extended anticoagulation decision is a risk–benefit trajectory, not a fixed duration—Sanfilippo framework, API-CAT reduced dose, EVE data.
  4. Apixaban and rivaroxaban are pragmatically equivalent for extended VTE prevention (RENOVE post-hoc).
  5. Concurrent antiplatelet therapy with prophylactic-dose apixaban doubles bleeding risk (AVERT post-hoc HR 2.02 for CRB)—justify every antiplatelet in a cancer patient on AC.
  6. At end of life, de-escalate to prophylactic-dose LMWH or DOAC, stop in the final days; bleeding clusters in the final week.

10.6 Case Synthesis

Ten Take-Home Points from “Turning the Tide”
  1. Cancer-associated VTE is the unifying topic: DVT from mass effect (Mr TYC), VTE in high-bleeding-risk tumours (Mr M), and VTE under drug-interaction pressure (Mr MT).
  2. DOACs are standard for most CAT (CARAVAGGIO, SELECT-D; confirmed by real-world VICTORIE: apixaban HR 0.67 for overall bleed vs LMWH, rivaroxaban HR 0.32 for ICH).
  3. Luminal upper-GI cancers remain an LMWH indication; apalutamide and other strong CYP3A4/P-gp modulators take DOACs off the table.
  4. Cancer distal DVT is not benign—RIETE 1-year mortality 30.7%; treat with therapeutic AC.
  5. Primary brain cancer is the surprise DOAC winner: DOACs reduce bleeding ~0.35× vs LMWH, despite (or because of) the high baseline ICH risk.
  6. Extended reduced-dose DOAC beyond 6 months is effective and safer (API-CAT apixaban 2.5 mg BID: non-inferior for recurrence, superior for bleeding).
  7. Apixaban and rivaroxaban are pragmatically equivalent for extended VTE prevention, subject to drug-interaction caveats (RENOVE post-hoc).
  8. Bleeding during AC is a dynamic risk, not a static score—ADAPT-BLEED (AUC 0.72) and cancer-specific inpatient RAMs (Martin et al.) outperform legacy tools.
  9. Concurrent antiplatelets double bleeding risk on prophylactic-dose apixaban (AVERT post-hoc HR 2.02) without efficacy benefit—justify every antiplatelet.
  10. At end of life, de-escalate to prophylactic dose and discontinue in the final days: 92% of late-phase bleeds occur in the last week (Noble et al.).

10.7 Key References

  1. Lee LH, Rojnuckarin P, Weitz JI. Turning the Tide: Key Cases and Conundrums in Thrombosis Medicine. Highlights of ASH Asia-Pacific 2026 [slide deck p.1–p.105].
  2. Brodin E, Young A, Langer F, et al. VICTORIE: comparative safety and effectiveness of DOACs vs LMWH in cancer-associated VTE in Europe. ASH 2025 Oral Communication #332 [slide p.13–p.22].
  3. Likvornik R, Rokach L, Shapira B, et al. Dynamic machine learning-based prediction of major bleeding in patients receiving anticoagulant therapy (ADAPT-BLEED). ASH 2025 Abstract #abs25-10073 [slide p.32–p.39].
  4. Martin KA, Wilkinson K, Sparks A, et al. Development and validation of a risk assessment model for inpatient venous thrombosis in people with cancer. ASH 2025 [slide p.43–p.51].
  5. Leader A, Wilcox JA, Zwicker JI. Management of venous thromboembolism in patients with brain tumours. Blood. 2024;144(17):1781–1790 [slide p.42, p.57, p.63].
  6. Galanaud JP, Monreal M, Bertoletti L, et al. Cancer-associated isolated distal DVT: RIETE registry. J Clin Oncol. 2024;42:529–537 [slide p.56].
  7. Mahé I, Agnelli G, Ay C, et al. Extended reduced-dose apixaban for cancer-associated VTE (API-CAT). N Engl J Med. 2025;392:1363–1373 [slide p.61, p.82–p.85].
  8. De Lucena M, et al. Extended reduced-dose versus full-dose DOAC in cancer VTE: meta-analysis. J Thromb Haemost. 2026;24:573–582 [slide p.62].
  9. McBane RD, Loprinzi CL, Ashrani A, et al. Apixaban 2.5 mg vs 5 mg BID for extended CAT treatment (EVE trial). J Thromb Haemost. 2024;22:1704–1714 [slide p.81].
  10. Couturaud F, Presles E, Sanchez O, et al. Extended reduced-dose versus full-dose DOAC (RENOVE). Lancet. 2025;405:725–735; Girard P et al., post-hoc analysis, ASH 2025 [slide p.87–p.95].
  11. Djulbegovic B, Anuskiewicz S, Mazur J, Coltoff A, Greenberg C. Mechanical thrombectomy and long-term VTE recurrence. ASH 2025 Abstract #332 [slide p.73–p.75].
  12. Abramowitz SD, Marko X, D’Souza D, et al. DEFIANCE trial design. Am Heart J. 2025;281:92–102 [slide p.76].
  13. Saad M, et al. Impact of concurrent antiplatelet use on apixaban thromboprophylaxis in cancer: AVERT post-hoc analysis. ASH 2025 Abstract #741 [slide p.96–p.101].
  14. Leblanc K, et al. DDIs between DOACs and androgen-receptor-axis-targeted therapies. Cancers. 2024;16(19):3336 [slide p.71].
  15. Noble S, et al. Management of anticoagulation at end of life in cancer-associated thrombosis. BMJ Support Palliat Care. 2019 [slide p.104].
  16. Houghton DE, et al. Bleeding risk of DOACs vs enoxaparin in GI and non-GI cancer VTE (prospective observational study) [slide p.12].
  17. Lyman GH, Carrier M, Ay C, et al. ASH 2021 guidelines: venous thromboembolism in patients with cancer. Blood Adv. 2021;5(4):927–974 [slide p.68].
  18. Carrier M, et al. Choosing DOAC or LMWH in cancer-associated thrombosis. Curr Oncol. 2021;28(6):5434–5451 [slide p.69].
  19. Sanfilippo KM, et al. Duration of anticoagulation following cancer-associated thrombosis. Br J Haematol. 2022 [slide p.79].
  20. Li A, et al. Bleeding in real-world cancer cohorts. Blood. 2024;144:2280 [slide p.52].