Isolated Dermal Arteriolar Thrombosis in a Pediatric Patient: A Rare Complication During Eltrombopag Therapy for Chronic Immune Thrombocytopenia

To the Editor: Eltrombopag, an oral thrombopoietin receptor agonist (TPO-RA), is increasingly used for the treatment of chronic immune thrombocytopenia (ITP) in adults and children because of its convenience and favorable safety profile. Although thrombotic events have been reported as a potential adverse effect of TPO-RAs, they are uncommon. In adults, the reported incidence is approximately 1%–5% 1-5. In contrast, thrombotic complications in pediatric patients remain rare, with only a limited number of cases reported 6-9. Furthermore, thrombotic events typically occur in the venous system, such as deep vein thrombosis (DVT), whereas arterial or peripheral microvascular thrombosis is uncommon. Here, we report a rare case of biopsy-proven dermal arteriolar thrombosis of the toes occurring during eltrombopag therapy in a pediatric patient with ITP in the absence of other thrombotic risk factors. The patient was a 13-year-old male with no significant past medical history. He was diagnosed with ITP at 11 years of age. Although he initially responded to intravenous immunoglobulin (IVIG) and prednisolone, the disease course was complicated by repeated relapses. Therefore, eltrombopag therapy was initiated 7 months after diagnosis at a starting dose of 12.5 mg/day and gradually increased to a maximum of 50 mg/day to maintain an adequate platelet count. After initiation, platelet count remained stable, generally above 50 × 109/L (Figure 1). Eleven months after starting eltrombopag, the patient developed pain at the tip of his right first toe, followed 1 week later by the appearance of livedo. The symptoms gradually worsened and eventually resulted in gait disturbance, prompting referral to our department. On admission, physical examination revealed intense erythema and spontaneous pain in both first toes. Livedo with subcutaneous induration was also observed on the right fifth toe (Figure 2a,b). Despite marked tenderness, capillary refill time remained less than 2 s, suggesting preserved peripheral circulation. No fever, joint swelling, or limitation of motion suggestive of infection or inflammatory arthritis was present. Laboratory evaluation revealed a white blood cell count of 5000/µL, hemoglobin of 14.2 g/dL, and a platelet count of 125 × 109/L, which was not significantly elevated compared with previous values (Figure 1). Coagulation parameters were within normal limits, including a fibrinogen level of 271 mg/dL and a D-dimer level of 0.8 µg/mL Autoimmune and vasculitis-related antibodies, including antinuclear antibodies, anti-DNA antibodies, antineutrophil cytoplasmic antibodies, and antiphospholipid antibodies, were negative. Lower limb ultrasonography revealed no evidence of deep vein thrombosis. A skin biopsy from the right first toe was performed (Figure 2c,d). Histopathological examination with hematoxylin and eosin staining demonstrated multiple dermal vessels occluded by fibrin thrombi with minimal inflammatory cell infiltration, findings inconsistent with vasculitis and rather suggestive of thrombotic occlusion of dermal arterioles. Based on the clinical course and pathological findings, infectious diseases, vasculitis, and connective tissue disorders were excluded, and eltrombopag-associated dermal arteriolar thrombosis was considered the most likely diagnosis. Eltrombopag was discontinued, and the patient was treated conservatively with non-steroidal anti-inflammatory drugs and topical tocopherol ointment. Notably, the erythema, livedo, and pain improved rapidly after discontinuation of eltrombopag, further supporting a possible association between the drug and the thrombotic event. Thrombotic complications associated with eltrombopag are considered rare, particularly in pediatric patients. Most previously reported pediatric cases involved additional risk factors such as trauma-related immobilization or obesity with oral contraceptive use 6, 10. In contrast, our patient developed thrombosis without such factors. Moreover, thrombosis related to eltrombopag or ITP occurs predominantly in the venous system, such as DVT, while the present case demonstrates that it may also present in the form of microvascular thrombosis. The relationship between eltrombopag and thrombosis remains complex. ITP itself has been associated with a paradoxical risk of thrombosis independent of TPO-RA therapy 11-13. Furthermore, thrombotic events related to eltrombopag appear to be extremely rare in other diseases such as aplastic anemia 14, 15. Recent meta-analyses in adult ITP reported no statistically significant increase in thromboembolic events associated with TPO-RA treatment compared with controls, although elevated risk ratios were observed 16. However, these analyses were limited to adult populations and cannot be directly extrapolated to pediatric patients. The exact mechanism of eltrombopag-associated thrombosis remains unclear. Patients with ITP have been reported to exhibit elevated levels of plasminogen activator inhibitor-1, resulting in impaired fibrinolysis and a prothrombotic state 17, 18. TPO-RAs may further enhance platelet activation and contribute to the formation of platelet-rich, fibrinolysis-resistant thrombi 19. A similar case of dermal microvascular thrombosis associated with eltrombopag has been described in an adult patient 20. In that report, a woman with chronic ITP developed livedoid skin lesions of the toes during long-term eltrombopag therapy, and biopsy confirmed microvascular thrombi. The clinical and pathological similarities to our case are notable. In conclusion, we report a rare case of biopsy-proven dermal arteriolar thrombosis during eltrombopag therapy in a pediatric patient with ITP without identifiable thrombotic risk factors. Clinicians should consider cutaneous microvascular thrombosis when patients receiving eltrombopag present with painful livedoid lesions of the fingers or toes, and early diagnostic evaluation, including skin biopsy, may be warranted. The authors thank Cosmin Mihail Florescu (Medical English Communications Center, University of Tsukuba) for his editorial assistance. Written consent was obtained from the patient's legal guardians. The authors declare no conflicts of interest.