Background: Immune thrombocytopenia (ITP) is a rare autoimmune disease, leading to bleeding and paradoxically to thrombosis. Eltrombopag and romiplostim are the two thrombopoietin receptor agonists (TPO-RAs) available in France and have been identified as a risk factor for arterial (AT) et venous thrombosis (VT) in ITP. The risk may be different between both drugs due to distinct pharmacological mechanisms. However, no direct comparison has been conducted to date. Based on data from the WHO pharmacovigilance database Vigibase®, a signal for a higher risk of thrombosis with eltrombopag compared to romiplostim has been exhibited in 2015, with a reporting odds ratio of 1.72 (95% CI: 1.47 – 2.02). This safety signal has not been confirmed so far. Therefore, the aim of this study was to compare the risk of thrombosis associated with eltrombopag versus romiplostim in patients with primary ITP.

Methods: A comparative cohort study emulating a target trial was designed in the FAITH cohort. This cohort is built with a previously validated algorithm in the French national health insurance system database (SNDS), which links socio-demographics, outpatient and hospitalization data for the entire French population (>67 million individuals), including diagnoses (using the International Classification of Diseases, 10th version – ICD-10) and outpatient drug dispensing in the whole country. Newly-diagnosed adults with primary ITP and exposed for the first time to a TPO-RA between 2011 and 2022 were selected. Secondary and prevalent ITP were excluded using an observation period ≥2 years. Patients with intracranial or gastro-intestinal bleeding between ITP diagnosis and TPO-RA initiation were excluded to avoid a channeling bias favoring romiplostim. Patients were assigned to the group corresponding to the first TPO-RA dispensed and were followed from TPO-RA initiation (index date) until the first hospitalization for thrombosis (identified using validated ICD-10 codes), TPO-RA discontinuation, death or after 1 year of follow-up. Randomization was emulated using the Inverse Probability of Treatment Weighting method, based on a propensity score computed with the following variables: age, sex, individual risk factors for AT and VT, year of TPO-RA initiation, duration of ITP and number of previous maintenance treatments received at index date. Cox proportional hazard regression models were computed with further time-dependent adjustment for the occurrence of events impacting the risk of thrombosis during follow-up: autoimmune hemolytic anemia (Evans syndrome), splenectomy, corticosteroids, intravenous immunoglobulin (IVIg), antiplatelet and anticoagulant drugs, oral contraceptives and estrogen replacement therapies. Secondary analyses were conducted by thrombosis type (venous and arterial) and by age subgroups (18-49, 50-74 and 75+ years old).

Results: Between 2011 and 2022, 3,710 adults with primary ITP exposed to a first TPO-RA were identified: 2,761 in the eltrombopag group and 949 in the romiplostim group. Patients in the romiplostim group were older, had more risk factors for thrombosis and a more severe ITP (based on IVIg use before TPO-RA initiation). In the weighted population, all covariates were well balanced.

The median follow-up time was 80 days (Q1-Q3: 42-152) in the eltrombopag group and 88 days (Q1-Q3: 42-198) in the romiplostim group. In the eltrombopag group, 74 hospitalizations for thrombosis (AT n= 29; VT n=45) were identified, compared to 18 (AT n=6; VT n=12) in the romiplostim group. The 1-year cumulative incidence was 5.1% (95% CI: 3.6-6.6) and 1.9% (95% CI: 1.0-2.7), respectively. The incidence of thrombosis at 6 months was similar between both groups. After 6 months, only 5 events occurred in the eltrombopag group while none occurred in the romiplostim group. The weighted, adjusted Hazard Ratio (HR) for eltrombopag versus romiplostim was 1.93 (95% CI: 1.11-3.34). Results were similar for VT (HR: 2.04; 95% CI: 1.04-4.03) and AT (HR: 1.99; 95% CI: 0.78-4.92) considered separately. In age subgroups, the risk difference was the highest in patients aged 75+ years (HR: 2.30; 95% CI: 1.03-5.14).

Conclusion: Although relying on few events, this cohort study suggests a higher risk of hospitalization for thrombosis in ITP patients treated for more than 6 months with eltrombopag than with romiplostim, particularly in the elderly.

This content is only available as a PDF.
2025
Sign in via your Institution