Effectiveness and safety of splenectomy or thrombopoietin receptor agonists as second-line treatments for patients with immune thrombocytopenia in a real-world setting

Abstract According to current guidelines, corticosteroids are well established as the first-line treatment for immune thrombocytopenia (ITP). For second-line therapy, splenectomy and thrombopoietin receptor agonists (TPO-RAs) are conditionally recommended. Splenectomy remains the only treatment known to induce long-term, treatment-free remission but is an invasive, irreversible procedure associated with lifelong risks such as post-splenectomy infections and thromboembolic events. Although the guidelines recommend tailoring second-line treatments to the characteristics of individual patients, this approach often lacks clarity in real-world practice. Therefore, we evaluated the comparative effectiveness and safety of splenectomy versus TPO-RAs as the second-line therapy in patients with ITP using a nationally representative dataset and applied statistical methods to minimize confounding in a real-world setting. This study used National Health Insurance claims database, covering ~98% of the population under a universal single-payer system. We included patients diagnosed with ITP (KCD-8 code D693) between 2002 and 2021 who (1) had a bone marrow exam within ±3 months of diagnosis, and (2) had no history of hematologic malignancies, aplastic anemia, or inherited bleeding disorders. From the eligible cohort, we identified two groups: those who underwent splenectomy (Group 1) and those treated with TPO-RAs without splenectomy (Group 2). Propensity score matching was performed between patients in Group 1 and Group 2 who had received corticosteroids and/or IVIG or anti-D immunoglobulins as first-line treatment. A total of 6,906 patients with ITP were identified. Among them, 777 patients (11.2%) underwent splenectomy (Group 1), and 100 patients (1.4%) received TPO-RAs without undergoing splenectomy (Group 2). Matching was based on age, sex, Charlson Comorbidity Index, and the interval between the date of ITP diagnosis and the initiation of splenectomy or TPO-RA treatment. Propensity score matching was conducted using a caliper width of 0.05. Post-matching balance was assessed by visualizing the distribution of propensity scores and by performing the Kolmogorov–Smirnov test, which yielded a K–S statistic of 0.25, a p-value < 0.001, and a standardized mean difference of 0.014, indicating acceptable covariate balance. After matching, 447 patients in Group 1 and 57 patients in Group 2 were included in the final analysis. Clinical outcomes were then compared between these matched groups. To evaluate effectiveness, we compared the time to next treatment between the two groups. Subsequent treatment was defined as the re-administration of corticosteroids, IVIG or anti-D immunoglobulins, TPO-RAs, or rituximab. In Group 2, retreatment with TPO-RAs was defined as a new prescription issued more than 6 months after the initial TPO-RA prescription date, based on insurance coverage criteria. Subsequent treatment was administered to 82.3% of patients in Group 1 and 61.4% of patients in Group 2. Event-free survival (EFS) was 5.03 months (range, 0–213.46) in Group 1 and 80.13 months (range, 0–230.1) in Group 2. In Group 1, most patients received the next treatment within 3 years after splenectomy, while in Group 2, it occurred within 3 months after the initiation of TPO-RA. During the first 24 months, the difference in median event-free survival was not statistically significant. However, from 24 months onward, Group 2 demonstrated significantly better EFS compared to Group 1. Pneumococcal infections were reported in 14 patients (3.1%) in Group 1 and none in Group 2. Neisseria meningitidis infections were not observed in either group. Haemophilus influenzae infections were reported in 7 (1.5%) patients in Group 1 and 1 patient (1.7%) in Group 2. Venous thromboembolism (VTE) was observed in 47 patients (10.5%) in Group 1 and 9 patients (15.7%) in Group 2. Arterial thromboembolism (ATE) was observed in 12 patients (2.6%) in Group 1 and 1 patient (1.7%) in Group 2. In this real-world analysis, within the first 24 months, no superior treatment between splenectomy and TPO-RA was identified. However, after 24 months, TPO-RA demonstrated better EFS. Regarding safety, splenectomy was linked to a higher risk of pneumococcal infections and ATE, while TPO-RAs were associated with an increased risk of VTE.