Immune thrombocytopenia (ITP) is an autoimmune disorder that manifests as peripheral thrombocytopenia (platelet count under 100,000/µL) resulting from increased platelet destruction and the inhibition of megakariopoiesis(1). Children with ITP have self-limited disease, with mostly favorable outcomes and spontaneous improvement. Adult patients, on the other hand, have a lower rate of remissions, approximately 40% by one year and 60% over three years(2). Up to 75% of them develop chronicity(3).
Current treatment options include steroids (high doses of dexamethasone or prednisone), immunoglobulin or a combination of both. Second-line therapy is primarily based on thrombopoietin receptor agonists or rituximab, while splenectomy is preferably delayed until at least one year after diagnosis. Additional second-line treatment is based on fostamatinib or immunosuppressive agents (azathioprine, cyclosporine, mycophenolate mofetil), but there are no guidelines to specify in which order the second-line therapy should be used(4,5).
However, there is a small group of refractory patients who do not respond to any treatment.
For this category of patients, current options include either continuing the previous treatment, initiating TPO-RA or rituximab, either alone or in combination with dexamethasone.
Alternative medications that have shown some success are low-dose prednisolone, dapsone, danazol and other immunosuppressive agents, but few data about their efficacy or toxicity with long-term use are available(6). Splenectomy is the last option, and the current guidelines recommend to delay it until over one year from ITP diagnosis.
Emerging strategies targeting pathological pathways such as phagocytosis inhibition targeting FCgR signaling (spleen tyrosine inhibitors, Bruton tyrosine kinase), inhibition of the neonatal Fc receptor (rozanolixizumab, nipocalimab, efgartigimob), inhibition of the classical complement pathway (sutimlimab) or inhibition of platelet desialylation (neuramidase-1 inhibitor), offer promising new alternative therapies for this category of ITP patients(7).
Definition of refractory immune thrombocytopenia
Refractory ITP was proposed by an international working group for patients who remain at a high risk for bleeding after failing previous treatments (at least two treatments), including splenectomy(8). The definition of response defined by Rodeghiero et al. is achieving a platelet count of 30,000/ml or doubling the baseline platelet count(8). Some of these patients may tolerate thrombocytopenia well (even low platelet count as 10,000/mL), leading to an almost normal quality of life, or they may choose to live with low platelet counts instead of undergoing potentially toxic treatments.
Recent studies have shown that the more lines of therapy a patient has failed, the more “refractory” he becomes to subsequent medical therapy(9). The major issue is associated with the fact that having a very low platelet count can increase the risk of serious bleeding.
Risk factors associated with increased incidence of bleeding are older age, comorbidities, need for antiplatelet agents or anticoagulation therapy, polypharmacy and refractoriness of disease. Patients with refractory ITP require vigilant care because they are at a serious risk of major bleeding, respond poorly to various treatments, develop worsening disease or medication-induced toxicities, they have a lower quality of life and higher hemorrhagic and infectious morbidity and mortality, especially in older age(2).
Biology of refractoriness
The effectiveness of the treatment depends on the initial trigger of the thrombocytopenia, which is multifactorial and targets components of adaptive immunity (T and B cells), as well as inflammatory factors (cytokines and autoantigens), along with impaired platelet production(10). Therefore, therapies targeting these aspects of the disease may be more efficient than administering a treatment that affects one or another. The duration of disease is relevant in terms of response to treatment. The chronic phase may be a result of B and T-cell memory response which is more difficult to resolve than the primary response that results in the induction of immune thrombocytopenia(11).
Another possible mechanism is antigen/epitope spread, generating new antiplatelet antibodies or the evolution to MDS. If ITP transitions from being primarily antibody-driven to T-cell driven, it becomes more challenging to treat(2).
In a small number of patients with refractory ITP, unresponsive to rituximab, TPO-RA or splenectomy, clonal expansion has been reported(2). Also, expansion of long-lived plasma cell in spleen in an enrich BAFF environment may contribute to the rituximab refractoriness(12).
For patients who have failed multiple therapies, it is important to reconsider the diagnosis first and perform a bone marrow evaluation, if not already done. Up to 20% of apparent cases of primary immune thrombocytopenia are secondary ITP(13,14). Secondary immune thrombocytopenia and myelodysplastic syndrome were the most common misdiagnoses associated with ITP(15).
Inherited thrombocytopenia, drug-induced thrombocytopenia and bone marrow failure with thrombocytopenia represented other common possibilities(2). Some of these entities may respond to ITP treatment, making them even more impossible to diagnose. Refractory patients should undergo extensive evaluation with bone marrow examination, aspirate and biopsy, cytogenetics and flow cytometry. Whole genome sequencing is a reasonable next step when bone marrow evaluation is normal(2). Even though all these examinations are negative, it remains very challenging to distinguish between refractory ITP and secondary thrombocytopenia.
The next step is reassessing the need for treatment, adequacy of prior therapies, along with the risk and benefits of further treatment. Approximately 20% of patients do not maintain a hemostatic platelet count after splenectomy or fail to respond to initial or subsequent therapies, but not all need further treatments as regarding platelet count/bleeding risk(16). Now, more than ever, the physician should also consider the long-term toxicity and the impact on patients’ quality of life, and consider discussing splenectomy as a secondary option.
Current treatment options
One of the most challenging questions regarding refractory ITP patients who require treatment after failing multiple therapies is how to approach their treatment. While the latest guidelines offer some guidance, there is still a need for significant research and development.
Approximately two-thirds of patients with ITP treated with rituximab as a second-line option show a response at one month, with a 40% response at one year, and a 30% response at the five-year follow-up(17). Compared with other second-line therapies, splenectomy offers a high response rate of 90% at one month, with 78% durable responses at ten years of follow-up(18). TPO-RA have a 65% response rate at one month, with 63% durable responses(9). They are all very good options, but rituximab has a favorable safety profile risk compared with splenectomy and is a time-limited therapy compared with TPO-RA.
Rituximab is an anti-CD20 monoclonal antibody that depletes CD20+ lymphocytes by triggering apoptosis through antibody dependent cell-mediated cytotoxicity and complement mediated lysis(19).
Rituximab therapy has the potential of inducing long-term responses in a subset of patients, though the long-term remission rates have generally been disappointing(20). It is preferred over long-term medication for patients with thrombotic risk, in those who require more sustained increases in platelet counts, or in patients with autoimmune disorders who require immunosuppression (rheumatoid arthritis, lupus erythematosus)(16). A number of different doses have been used with no superiority regarding the response rate, but four weekly infusions of 375 mg/m2, or alternative dosing schedule of 1000 mg on day 1 and 100 mg weekly for four weeks seem to associate response rates over 50%(4).
The adverse effects of rituximab include infusion reactions, serum sickness and cardiac arrythmias. Reactivation of hepatitis B is more frequent, and active hepatitis B is a contraindication for rituximab. Reactivation of latent John Cunningham polyomavirus (or JC virus), causing progressive multifocal leukoencephalopathy, has been very rarely reported.
Rituximab treatment alter the immunological response, resulting in a 6-12-month immunosuppression which may diminish the vaccine response(16).
Response rates to rituximab vary from 60% to 80%(21), and predictors of success seem to be the shorter duration of disease, age under 40 years old and female gender(22); on the other hand, patients with an increased number of failed previous treatments and longer duration of disease have a lower probability of response(22). Previous splenectomy does not affect the response rate, but an earlier relapse tendency was observed(23). Concomitant use of corticosteroids may improve the response rates(23).
TPO-RA (avatrombopag, eltrombopag, romiplostim) have provided excellent responses (above 60%) in both splenectomized and non-splenectomized patients(24). They increase platelet production by binding to the TPO receptor, causing conformational change and the activation of JAK/STAT5 pathway, resulting in increased megakaryocyte progenitor proliferation and increased platelet production(25). Although their end result is the same, each drug acts differently on the TPO receptor, by activating different signaling pathways in megakaryoctyes. Romiplostim is a peptide that binds directly and competitively at the TPO binding site, and eltrombopag is a small molecule which binds a transmembrane site(25). Eltrombopag acts earlier in the pathway and stimulates MK precursor cells and MK differentiation, while romiplostim stimulates mostly mature precursors(25). In addition to increasing platelet counts, they seem to have some immunomodulatory effects. TPO-RA seems to increase regulatory T and B cells effects, mediated by TGF-B, which is a major cytokine involved in Treg cell development, found in high levels in MK and platelets(25). These differences may explain variations of response in ITP patients(26). Romiplostim starting dose is 1 µg/kg/week subcutaneously, with dose adjustments up to 10 µg/kg/week, according to platelet responses(4). Eltrombopag starting initial dose is 25 or 50 mg/day, depending on patient’s age or hepatic impairment, up to a maximum of 75 mg/day(4). Although they provide a very good response, they are considered a maintenance therapy, with platelet counts usually declining following cessation of the therapy. However, in approximately one-third of patients, the treatment can be discontinued successfully. Among the adverse reactions, the most important are bone marrow reticulin fibrosis, which appears to be reversible after the cessation of treatment, mild liver test abnormalities (eltrombopag), and arterial or venous thrombosis(11,16). Avatrombopag, like eltrombopag, increases platelet production after binding the transmembrane region of TPO receptor and activating signal transduction pathways(27). It is administered orally, with a starting dose of 20 mg/day up to 40 mg/day, depending on platelet count; it has no interaction with food and it does not affect the liver function(28). It offers a rapid platelet response (66% in day 8), which is maintained for over 12 months, as reported in a recent study(29).
Fostamatinib is a spleen tyrosine kinase inhibitor with indication for adults with chronic ITP who had insufficient response to previous treatments. With a starting dose of 100 mg twice daily up to 150 mg in nonresponders, an overall response rate of 43% was achieved, but with a median stable response (platelet count over 50 x109/L) of 18%, maintained on the long-term treatment(30). More frequent adverse reactions are arterial hypertension and diarrhea, which require dose reduction(30).
Mycophenolate mofetil showed (in retrospective studies) an approximate 50% response rate in adults with primary ITP, although there are no randomized clinical trials (RCT) of MMF in ITP. It is an oral medication with a starting dose of 500 mg twice a day, with a response rate seen after 6-8 weeks. The reported side effects are represented by headaches, gastrointestinal toxicity, liver function test abnormalities and increased risk of infections(6).
Among immunosuppressive agents, danazol and dapsone may be useful in corticoid dependent patients, in those with contraindication to splenectomy or if other agents are not available.
Danazol is a synthetic androgen that antagonizes estrogens, and it is postulated to decrease Fc receptor numbers on phagocytic cells(31). It has shown good response rates either alone or in combination with dexamethasone, but it seems to be more effective, with a shorter response rate, when combined with rhTPO(32). The usual dose is 200 to 800 mg per day, with a response rate ranging from 23.8% to 57.9% at one month, and with a durable response ranging from 9% to 96%(16). Androgenic effects, elevated liver function tests, weight gain, acne and rash are the most common adverse reactions(31).
Dapsone is an antibiotic indicated for infections, skin conditions and autoimmune diseases(31). It has anti-inflammatory action by inhibiting neutrophil myeloperoxidase and acts as a potent oxidant by affecting glucose-6-phosphate-dehydrogenase (G6PD); hemolyzing red blood cells saturate the phagocytic capacity of macrophages, with sparing platelets from destruction(33). Administered orally, with a starting dose of 50 to 100 mg daily, it has a variable response rate of 36-63% reported at one month and a durable response varying from 0% to 55%(34). It is well tolerated, but nausea and mild hemolysis can occur(31).
Splenectomy was the gold standard for many decades, mainly because of the high rates of durable remission. The current recommendations of the ITP International Working Group and ASH guidelines are to delay splenectomy at least for one year after the initial disease onset and to offer it to patients who relapse or do not achieve response to steroids. Ideally, all medical treatments should be offered before surgery(5,16). With an initial response rate up to 90% and a median response rate of 78% at ten years and 68% at 20 years, splenectomy is the only option associated with long-term treatment-free remissions(18). Although not well defined, it seems that patients over the age of 60 have higher relapse rates(35). Indium-labeled autologous platelet scanning is a useful assessment in confirming that the spleen is the main site of platelet destruction(36). The disadvantages of splenectomy include surgical complications, increased risk of post splenectomy infection and thrombosis. Postoperative thromboprophylaxis for patients with platelet count over 30-50 x109/L, appropriate vaccination against Streptococcus pneumoniae, Neisseria meningitidis and Haemophilus influenzae type B provided two weeks before surgery and antibiotic prophylaxis are recommended for ITP patients who undergo splenectomy(5,16).
Management of patients experiencing treatment refractoriness
The management of patients who failed multiple therapies is a real challenge. The major issues facing these patients include significant morbidity and mortality, as well as lack of data in this field. The current guidelines suggest some recommendations, but relevant studies are necessary to develop an algorithm.
Before TPO-RA, multiple combination of immunosuppressant treatments have been used, with some results in small clinical trials, but the actual tendency is to limit immunosuppressive treatment because of their long-term adverse reactions(2).
For patients previously treated with TPO-RA, new data have shown that switching from one TPO-RA to another one has a positive effect on response and tolerability, being effective in 50% to 80% of patients, with the resolution of platelet count fluctuations and adverse reactions(37). Another study revealed that switching from romiplostim to eltrombopag led to a 100% response rate, but from eltrombopag to romiplostim led to an only 60% success rate if the switch was made as a result of the inefficacy of the first treatment. TPO-RA response is corelated with the TPO level(38). At a thrombopoietin level over 400 pg/ml (normal range: below 100), less than 10% of patients respond to TPO-RA(39). For these patients, TPO-RA should be stopped and another treatment option must be taken into consideration(39).
Combining immunosuppressant treatments with TPO-RA shows promising responses, but serious randomized controlled trials are necessary(2,31). TPO-RA was studied in association with immunosuppressive agents (cyclosporine, mycophenolate mofetil), FcR blocking agents (i.v. immunoglobulins, fostamatinib) or both. The complete response was obtained in over 50% of multirefractory ITP patients(40).
Promising results (72% of patients with chronic ITP) were offered by a combination of TPO-RA with i.v. immunoglobulins and an immunosuppressant agent (cyclosporine or mycophenolate mofetil). A study highlighted the importance of using combination of agents with different mechanisms of action in order to achieve the best results(40).
Different combinations with rituximab were also investigated, but so far without yielding encouraging outcomes. Two controlled studies that compared dexamethasone alone or in combination with rituximab demonstrated superior response at six and 12 months when adding rituximab, but with no difference regarding long-term follow-up and with higher adverse reaction and toxicity reported in rituximab arms(41,42).
Another trial from China investigated on 105 patients with relapsed/ refractory ITP the efficacity of four low-dose rituximab infusions with recombinant human TPO (300 mg) for 14 days versus rituximab alone. Although an increase in platelets count was obtained along with reduced bleeding, the study didn’t reveal a difference in sustained response. Also, the rhTPO is available only in a few countries beside China(43).
Alemtuzumab showed efficacy only associated with rituximab, but this is not recommended due to the dual immunosuppression(44).
In patients who undergo and responded to splenectomy but relapsed after more than one year, an accessory spleen must be investigated and resected, if found(45).
Hematopoietic stem cell transplant (HSCT) is reserved only for patients with severe chronic nonresponsive ITP and bleeding complications unresponsive to other treatments. This procedure carries a very high risk, therefore it is recommended in selected cases.
Patients refractory to rituximab and TPO-RA will benefit without doubt from using a combination of both treatments.
On demand therapy is the only treatment used at the time of or in anticipation of high-risk bleeding situations. In some cases, the lowest effective dose of corticosteroids is maintained for weeks or months. Dapsone could be a useful alternative to spare corticosteroids. Also, danazol could be a useful alternative until a decision on splenectomy is made(46). Danazol and all-trans retinoid acid showed a sustained response in 63% of 45 patients compared with 26% of 48 patients with danazol as monotherapy(47).
For the patients who remain refractory to all medications, with acute bleeding episodes, clinical trials or novel targets are recommended.
New advanced therapeutic molecules target different pathways involved in ITP pathogenesis, from peripheral destruction of platelets to inappropriate megakaryopoiesis. Inhibition of phagocytosis targeting FCgR signaling (spleen tyrosine inhibitors, Bruton tyrosine kinase), inhibition of the neonatal Fc receptor (rozanolixizumab, nipocalimab, efgartigimob), inhibition of the classical complement pathway (sutimlimab), or inhibition of platelet desialylation (neuramidase-1 inhibitor) will modify the management of ITP in the near future. Some therapies, currently under investigation in clinical trials, seem to offer promising results. These include new anti-CD20 monoclonal antibodies (veltuzumab, obinutuzumab), proteasome inhibitors or BAFF targeting therapies (belimumab, blisibimob) and anti-CD40 monoclonal antibodies (which target TFH implicated in ITP pathogenesis via CD40L/CD40 axis and IL 21 production). Additional therapy may consist of low doses of IL-2 or histone deacetylase (clidamicine) in order to restore the Treg function(7,48).
In the era of multidrug ITP therapies, a minority of patients still do not show response after several treatments. Consequentially, these patients demonstrate severe disease with high morbidity and are at a higher risk of death. In the absence of specific recommendation of guidelines, an individual approach based on combination of agents appears to be the best way of improving the outcome of patients with multirefractory ITP.
Using a TPO agent that stimulates megakaryopoiesis in combination with medication that ameliorate the immune attack and inhibit platelet destruction seems to be the right answer.
Significant work is still needed regarding what combination therapy, which agents to use, at what dose, how long to give them, and to which patients.
New molecules with novel mechanisms of action will soon be available and will alter the management and prognosis of ITP. However, the need for biomarkers to help clinicians in diagnosis and tailoring treatments remains crucial.
Conflict of interest: none declared.
Financial support: none declared.
This work is permanently accessible online free of charge and published under the CC-BY licence.