Successful use of eltrombopag without splenectomy in refractory HIV-related immune reconstitution thrombocytopenia

We report the successful use of eltrombopag, a novel thrombopoietin (TPO) receptor agonist, in a case of refractory HIV-related immune reconstitution thrombocytopenic purpura (ITP). A 47-year-old male carpet layer presented with recurrent bacterial pneumonia, weight loss and subcentimeter lymphadenopathy. Platelet count was low at 93 × 109/l (normal range >150 × 109/l), and the remainder of the full blood examination including hemoglobin and leucocyte counts were normal. HIV infection was confirmed by Western Blot assay. HIV-1 viral load was greater than 100 000 copies/ml and initial CD4 cell count was 123 × 106/l. Sulphamethoxazole/trimethoprim was commenced as Pneumocystis jirovecii pneumonia prophylaxis. Two weeks later Atripla (efavirenz/tenofovir/emtricitabine; Gilead Sciences, Foster City, California, USA) was commenced. Two weeks following the commencement of antiretroviral therapy, he developed petechiae and periorbital purpura and platelet count had dropped to 10 × 109/l. CD4 cell count was 250 × 106/l. Sulphametoxazole/trimethoprim was ceased and prednisolone 1 mg/kg orally (p.o.) daily was commenced. After 2 weeks of prednisolone, platelet count moderately improved to 20 × 109/l but spontaneously dropped to 14 × 109/l upon re-admission to hospital with sweats and headache. Intravenous immunoglobulin (IVIg) (1 g/kg) was given due to concern of intracranial haemorrhage, and platelet count promptly increased to 119 × 109/l but again decreased to 14 × 109/l over 2 weeks despite ongoing prednisolone (1 mg/kg/day). Bone marrow biopsy showed adequate trilineage hematopoiesis and increased megakaryopoiesis in keeping with a peripheral destructive aetiology for the thrombocytopenia. Cell surface makers as assessed by flow cytometry of the bone marrow aspirate showed no evidence of a clonal lymphoproliferative process, and antiphospholipid antibodies were not found. Azathioprine 100 mg/day was then added. Over the next 3 months, platelet count fluctuated between 10–20 upon repeated failed attempt at corticosteroid weaning, and the patient was again readmitted with rectal bleeding when prednisolone was weaned to 35 mg/day, with a platelet count of 10 × 109/l, and required one dose of IVIg (1 mg/kg). Azathioprine was ceased, and sulphamethoxazole/trimethoprim was recommenced with the re-escalation of prednisolone to 50 mg p.o. daily in combination with hydroxychloroquine 200 mg twice daily. At this point, HIV viral load was undetectable (<50 × 106 copies/ml) and CD4 cell count 150 × 106/l approximately 3 months after the commencement of Atripla. Atripla was ceased on the premise that it may have contributed to the thrombocytopenia, and an alternative antiretroviral regimen was commenced with abacavir–lamivudine and ritonovir-boosted darunavir. Despite all these therapeutic manoeuvres, corticosteroids could not be completely weaned, and platelet count fluctuated between 10–20 × 109/l over the following 3 months. The patient was unable to return to work because of its physical nature. Splenectomy was considered but was avoided due to concern regarding potential infectious complications [1]. It was then decided to trial eltrombopag 50 mg p.o. daily as an alternative treatment for refractory HIV-related thrombocytopenia. After 1 week, platelet count rose to 87 × 109/l. Prednisolone was weaned and ceased after 8 weeks and he has returned to work. Currently, 6 months since commencement of eltrombopag, platelet count is stable between 90–126 × 109/l, HIV viral load is undetectable, and CD4 cell count is more than 400 × 106/l. He has been subsequently changed back to Atripla for antiretroviral therapy without harmful effect. Thrombocytopenia has been associated with HIV since the beginning of the AIDS epidemic [2,3] and occurred relatively frequently and early in the period before highly active antiretroviral therapy (HAART) [4,5]. Shortly after they became available, antiretrovirals were shown to be successful in treating ITP in a randomized trial of zidovudine monotherapy [6]. Steroids and splenectomy were also used with some success, but the use of HAART has reduced the severity and frequency of HIV-related ITP to the point where splenectomy is hardly ever required [1,7]. Paradoxically, the ITP in this case deteriorated 2 weeks after commencement of HAART in the setting of improved CD4+ cell count, and became progressively refractory to corticosteroids, azathioprine and hydroxychloroquine despite the eventual achievement of undetectable viral load and normal CD4+ cell count. Unlike most HIV-associated ITP that improves with HAART, this clinical deterioration during immune recovery is suggestive of an immune reconstitution inflammatory syndrome (IRIS). Here, restored immunity results in the emergence and/or dysregulation of immunological response to various infectious and noninfectious antigenic stimuli, including that from autoantigens as in this case. Immune reconstitution-related ITP in patients with HIV has not been widely described, except for one previous case report that demonstrated the successful use of corticosteroids [8]. However, the use of corticosteroids, which are potent CD4+ T cells suppressors, needs to be weighed carefully against the need to prevent opportunistic infections. Eltrombopag is a small molecule, oral TPO-receptor agonist that interacts with the TPO-receptor leading to JAK/STAT pathway signalling in megakaryocytes [9]. This leads to stimulation of bone-marrow platelet production that keeps up with the rate of destruction in ITP and indeed, this drug is Food and Drug Administration-approved for the treatment of chronic ITP. Response is prompt and is usually within 2 weeks of treatment, as was seen in our patient. Eltrombopag is well tolerated and its side-effect profile is comparable to that of placebo in previous studies, with no increase in thromboembolic phenomenon [10]. Increased bone marrow reticulin has been reported with long-term use of TPO-receptor agonists, including romiplostim; however, this was found to be reversible upon treatment cessation [11]. Although TPO-receptor agonists have the advantage of not being immunosuppressants, the use of eltrombopag in HIV is still being investigated. Indeed, a recent pharmacokinetic study has shown that lopinovir–ritonovir decreases the plasma concentration of eltrombopag but not vice versa, implying that although the dose of eltrombopag may need to be adjusted upon the coadministration antiretrovirals, it does not decrease the efficacy of the latter [12]. Currently, no standard treatment exists for the treatment of IRIS in patients with HIV, and management varies depending on patient's clinical presentation, the ability to identify and eradicate the underlying infective pathogens/antigenic stimulants, and the need to dampen the immune response. In IRIS-related ITP, immunosuppressive therapy is vexed by the increasing risk of opportunistic infections. Here, we have demonstrated that eltrombopag is a useful option, even in refractory cases, especially when splenectomy is not preferred. Eltrombopag should be further investigated as a rescue therapy in HIV-related ITP. In addition, our case likely represents a second report of ITP presenting as an immune reconstitution syndrome, and highlights both the lack of current treatment guidelines and the need for further studies in this area. Acknowledgements Conflicts of interest There are no conflicts of interest.