AFFILIATIONS
1 Johns Hopkins University, Baltimore, Maryland, United States
2 Johns Hopkins University, Baltimore, Maryland, United States
3 Johns Hopkins University, Baltimore, Maryland, United States
4 Johns Hopkins University, Baltimore, Maryland, United States
5 Johns Hopkins University, Baltimore, Maryland, United States
ABSTRACT
Background: Thrombopoietin receptor agonists (TPO-RAs) are widely used to treat immune thrombocytopenia (ITP). Clinical trials have reported a slightly increased risk of thrombotic events with TPO-RAs, although absolute event rates were low. However, real-world evidence remains inconsistent, and it is unclear whether any observed increase in thrombotic risk is attributable to TPO-RA exposure, the intrinsic prothrombotic state associated with chronic ITP, or coexisting comorbidities. Furthermore, thrombotic events are often reported as composite outcomes, with arterial and venous events rarely analyzed separately. Using a large real-world database, we evaluated the association between TPO-RA exposure and the risks of arterial and venous thrombosis in patients with ITP and identified risk factors associated with these events.
Methods: We conducted a retrospective cohort study of 20,760 adults with ITP identified in TriNetX between January 1, 2020, and December 31, 2024. ITP was defined by an ICD-10 diagnosis code and a documented platelet count <100,000/µL at any time. Patients were classified as ever or never exposed to a TPO-RA. Arterial and venous thrombotic events were identified using ICD-10 codes. We compared crude thrombosis prevalence and incidence rates per 100 person-years between TPO-RA–exposed and unexposed patients. Multivariable logistic regression estimated adjusted odds ratios (aORs) for overall, arterial, and venous thrombosis, controlling for age, sex, race, recent hospitalization, and relevant cardiovascular, metabolic, pulmonary, renal, malignant, autoimmune, hormonal, and smoking-related risk factors. Cox proportional hazards models identified risk factors among TPO-RA-exposed patients. The proportional hazards assumption was assessed using scaled Schoenfeld residuals; because hypertension violated the assumption, it was included as a stratification variable.
Results: Among 20,760 patients, 3,017 (14.5%) were exposed to a TPO-RA. Follow-up totaled 3,097.26 person-years in the TPO-RA-exposed group and 69,303.44 person-years in the unexposed group, with mean observation periods of 1.0 and 3.9 years, respectively. The prevalence of overall, arterial, and venous thrombosis was 24.74%, 20.96%, and 7.08%, respectively. Incidence rates among exposed versus unexposed patients were 54.92 versus 50.64 per 100 person-years for overall thrombosis (IRR, 1.08; 95% CI, 1.03–1.14), 38.68 versus 40.68 for arterial thrombosis (IRR, 0.95; 95% CI, 0.90–1.01), and 16.34 versus 9.97 for venous thrombosis (IRR, 1.64; 95% CI, 1.50–1.79). After adjustment, TPO-RA exposure was associated with venous thrombosis (aOR, 1.20; 95% CI, 1.03–1.39; P=0.021), but not arterial thrombosis (aOR, 0.96; 95% CI, 0.85–1.08; P=0.512) or overall thrombosis (aOR, 1.05; 95% CI, 0.94–1.17; P=0.359). Age ≥65 years (aOR, 1.15; 95% CI, 1.004–1.309), antiphospholipid syndrome (aOR, 4.83; 95% CI, 3.51–6.65), atrial fibrillation (aOR, 2.09; 95% CI, 1.83–2.39), cancer (aOR, 1.40; 95% CI, 1.16–1.68), and recent hospitalization (aOR, 1.30; 95% CI, 1.16–1.46) were associated with venous thrombosis. Although TPO-RA exposure was not associated with arterial thrombosis, established risk factors, including ischemic heart disease (aOR, 5.60; 95% CI, 5.09–6.15), age ≥65 years, male sex, diabetes, hypertension, smoking, hyperlipidemia, antiphospholipid syndrome, cancer, and peripheral vascular disease, were associated with arterial events.
Conclusions: Among patients with ITP, TPO-RA exposure was associated with an increased risk of venous thrombosis but not arterial thrombosis. The apparent arterial signal in crude prevalence analyses disappeared after accounting for the substantially shorter follow-up among exposed patients, highlighting the importance of incorporating exposure time when evaluating thrombotic outcomes. Together, the incidence and adjusted analyses suggest that TPO-RA–associated thrombotic risk is more likely to involve venous than arterial events and is further modified by established thrombotic risk factors. Causal inference is limited by the observational design, unequal follow-up, diagnosis code–based outcome ascertainment, multiple comparisons, potential exposure misclassification, and residual confounding, including confounding by indication. Nevertheless, our findings support targeted assessment of venous thromboembolism risk before and during TPO-RA treatment rather than considering thrombosis as a single composite complication.
DISCLOSURES
Shing Chao, MD: No relevant conflicts of interest to declare.