Impact of thrombocytopenia on hemorrhage and recurrent thrombosis in patients with cancer associated thrombosis, a time varying analysis

Abstract

Abstract IntroductionBoth venous thromboembolism (VTE) and thrombocytopenia are frequent complications in patients with cancer. Managing cancer associated thrombosis (CAT) in the setting of thrombocytopenia presents a significant clinical challenge, as it requires balancing the risk of thrombosis with the risk of bleeding. There is limited data on how fluctuations in platelet levels over time influence the risks of recurrent thrombosis and bleeding. MethodsThis is a retrospective, single-center study of adults with active solid malignancies and pulmonary embolism (PE) and lower extremity deep vein thrombosis (DVT). Platelet counts were extracted electronically over a 12-month period after index VTE. Thrombocytopenia was defined as mild (platelet count <100 × 10³/μL), moderate (platelet count <75 × 10³/μL), or severe (platelet count <50 × 10³/μL), at baseline (within 15 days of the index VTE diagnosis) and following index VTE. Outcomes were manually extracted including major bleeding (primary); and clinically relevant bleeding (CRB) (composite of major bleeding and clinically relevant non-major bleeding) and subsequent venous or arterial thrombosis (secondary). Demographics, cancer site and stage, cancer treatments, anticoagulation therapy, and relevant medical history were collected through electronic and manual data extraction. Robust Poisson regression was used to estimate the adjusted risk ratio (aRR) and 95% confidence interval (CI) for mild baseline thrombocytopenia and each of the outcomes of interest. Time-varying Cox proportional hazards models were used to estimate the adjusted hazard ratio (aHR) and 95% CI to assess association between dynamic thrombocytopenia following index VTE and outcomes. Models adjusted for age, sex, cancer type and stage, chemotherapy, anticoagulation, antiplatelets, renal disease, liver disease. Each outcome model was also adjusted for history of relevant outcome. ResultsOf 285 eligible patients, mean age was 67.9 ± 11.4 years. Over two-thirds (69.6%) of patients had a PE alone, 23.1% lower extremity DVT alone and 7.3% both. Most patients had metastatic disease (67.6%) and had received chemotherapy (68.4%). The most common types of cancer were gastrointestinal (35.8%), lung (19.0%), and genitourinary (14.1%). Most patients (89.8%) received anticoagulation following their index VTE, including apixaban (51.4%), enoxaparin (29.0%) and rivaroxaban (8.5%). Baseline mild thrombocytopenia was present in 9.0% of patients. Mild baseline thrombocytopenia was not significantly associated with major bleeding (aRR: 0.87, 95% CI: 0.45–1.69), clinically relevant bleeding (aRR: 1.01, 95% CI: 0.76–1.32), or venous thrombosis (aRR: 1.15, 95% CI: 0.66–1.99). None of the patients who experienced arterial thrombosis had baseline thrombocytopenia. The cumulative incidence of major bleeding at one year following index VTE was 12.0% (95% CI: 8.4–16.3). When examining thrombocytopenia as a time-varying exposure, major bleeding was significantly increased with mild (aHR: 11.50, 95% CI: 5.25–25.17), moderate (aHR: 8.58, 95% CI: 3.63–20.24), and severe (aHR: 25.76, 95% CI: 8.78–75.55) thrombocytopenia. The cumulative incidence of CRB at one year was 36.5% (95% CI: 30.9–42.4). Risk for CRB were significantly increased for mild (aHR: 5.58, 95% CI: 2.75–11.32), moderate (aHR: 3.59, 95% CI: 1.42–9.08), and severe (aHR: 3.51, 95% CI: 1.10–11.22) thrombocytopenia. The cumulative incidence of venous and arterial thrombosis at one year following index VTE was 11.5% (95% CI: 7.9–16.1) and 7.7% (95% CI: 4.9–11.5) respectively. The hazards for venous thrombosis were significantly increased and similar for all levels of thrombocytopenia (mild: aHR: 33.52, 95% CI: 7.58–148.2; moderate: aHR: 41.13, 95% CI: 7.10–238.2; severe: aHR: 40.14, 95% CI: 6.28–256.5). There were no significant associations between arterial thrombosis and time-varying mild (aHR: 4.57, 95% CI: 0.63–33.01), moderate (aHR: 2.40, 95% CI: 0.19–30.27), or severe (aHR: 8.35, 95% CI: 0.71–97.86) thrombocytopenia. Conclusions Thrombocytopenia is a key risk factor of both major bleeding and VTE in CAT. The severity of thrombocytopenia directly parallels the risk of major bleeding and thrombotic recurrence, yet even mild reductions in platelet count markedly elevate these risks. Incorporating thrombocytopenia into risk-stratification tools and tailoring anticoagulation strategies accordingly will be essential for safely managing this frequent clinical challenge.