Second-line and subsequent therapies for immune thrombocytopenia (ITP) in adults

INTRODUCTION — Therapy for immune thrombocytopenia (ITP) differs for different patients: some do not require any treatment; others have a spontaneous remission or respond to first-line therapy with glucocorticoids; and others continue to have severe thrombocytopenia necessitating additional therapy.

Here we discuss our approach to management of ITP for adults who require additional therapy beyond glucocorticoids (second-line and subsequent therapies). The clinical manifestations, diagnosis, and initial treatment of adults and children presenting with ITP are discussed in detail separately.

●Adults

•Diagnosis – (See "Immune thrombocytopenia (ITP) in adults: Clinical manifestations and diagnosis".)

•Initial treatment – (See "Initial treatment of immune thrombocytopenia (ITP) in adults".)

●Children

•Diagnosis – (See "Immune thrombocytopenia (ITP) in children: Clinical features and diagnosis".)

•Initial treatment – (See "Immune thrombocytopenia (ITP) in children: Initial management".)

•Chronic disease – (See "Immune thrombocytopenia (ITP) in children: Management of chronic disease".)

●Pregnancy – (See "Thrombocytopenia in pregnancy", section on 'Immune thrombocytopenia (ITP)'.)

INITIAL CONSIDERATIONS

Treatment of bleeding — Treatment of patients with acute, severe bleeding and patients who require a higher platelet count for surgery or obstetric indications is discussed separately. (See "Initial treatment of immune thrombocytopenia (ITP) in adults", section on 'Critical bleeding' and "Initial treatment of immune thrombocytopenia (ITP) in adults", section on 'Surgery or delivery'.)

Indications for second-line therapy — The goal of ITP therapy is to provide a safe platelet count to prevent clinically important bleeding, not to normalize the platelet count [1]. Second-line therapy is used when first-line therapy (eg, glucocorticoids) does not raise the platelet count to a safe level or if tapering first-line therapy results in a decrease of the platelet count below a safe level (algorithm 1). (See "Initial treatment of immune thrombocytopenia (ITP) in adults", section on 'Whom to treat (indications for therapy)'.)

We often initiate second-line therapy at a slightly lower platelet count than used for initial therapy (ie, <20,000/microL rather than 30,000/microL that is used for initial therapy), especially if the platelet count <20,000/microL is persistent, recurrent, or associated with bleeding symptoms (mucosal purpura or more serious bleeding). A 2019 guideline on ITP from the American Society of Hematology (ASH) considers second-line therapy to be appropriate for these individuals after ≥3 months from diagnosis [1]. However, it is reasonable to start second-line therapy before three months in patients who do not have a response to first-line therapy or who experience a relapse after first-line therapy is tapered. The ASH guideline and an associated consensus report emphasize that therapy should minimize toxicity and optimize the patient's quality of life [1,2].

The importance of proceeding to second-line therapy is greater for patients with a higher risk of bleeding due to other comorbidities (eg, hypertension, renal insufficiency), the need for concomitant antiplatelet agents or anticoagulant medications, or athletic activities or careers with high risk for trauma. For patients for whom bleeding is less of a concern, second-line therapy may be deferred even at lower platelet counts.

COVID-19 vaccination — We sometimes delay coronavirus disease 2019 (COVID-19) vaccination for individuals who are in the midst of an active ITP flare, due to the possible exacerbation of thrombocytopenia. However individuals with chronic ITP do not need to delay vaccination, and COVID-19 vaccination should not be withheld for individuals who have received (or are receiving) ITP therapy and are not in the midst of a flare. (See "Initial treatment of immune thrombocytopenia (ITP) in adults", section on 'COVID-19 vaccination'.)

While COVID-19 vaccination should not be withheld for individuals who have received immunosuppressive therapies, patients and clinicians should be informed about the possible reduced vaccine response (especially after rituximab) and the importance of taking other precautions against COVID-19. Additional considerations related to COVID-19 vaccination in individuals receiving immunosuppressive therapies are presented separately. Patients and providers are advised to review the latest recommendations with respect to booster vaccine doses in patients who are receiving immunosuppressive therapy. (See "COVID-19: Vaccines", section on 'Immunocompromised individuals'.)

Caveats before proceeding to second-line therapy — Caveats are summarized in the figure (algorithm 2) and include the following:

●Diagnostic confirmation – Before proceeding to a second-line agent, it is especially important to confirm the diagnosis of ITP by eliminating other possibilities such as an inherited thrombocytopenia, myelodysplastic syndrome (MDS), or drug-induced thrombocytopenia. In one series of 295 individuals who were initially diagnosed as having ITP, 36 (12 percent) were subsequently reclassified as having a different diagnosis requiring a different management approach [3]. Review of the patient and family history and the complete blood count (CBC) and blood smear for platelet and other cell morphologies are especially important. (See "Immune thrombocytopenia (ITP) in adults: Clinical manifestations and diagnosis", section on 'Differential diagnosis'.)

●Possible secondary ITP – It is important to review possible causes of secondary ITP that might respond to treatment of the underlying condition, such as Helicobacter pylori infection, HIV infection, hepatitis C virus (HCV) infection, or chronic lymphocytic leukemia (CLL) [1]. (See "Immune thrombocytopenia (ITP) in adults: Clinical manifestations and diagnosis", section on 'Inciting events'.)

●Preparation for immunosuppression – Several second-line and third-line therapies are immunosuppressive, including splenectomy, rituximab, and other immunosuppressive agents. Even if these are not used right away, they may be used in the future. It is prudent to ensure that all regular vaccinations (including coronavirus disease 2019 [COVID-19] vaccination) are updated, and it may be reasonable to provide pre-splenectomy vaccinations in case splenectomy or rituximab is used. Individuals who may receive rituximab should also be screened for active or latent hepatitis B virus (HBV) infection and treated appropriately. (See 'Immunizations' below and "Elective (diagnostic or therapeutic) splenectomy", section on 'Vaccinations' and "Hepatitis B virus reactivation associated with immunosuppressive therapy".)

Immunizations — Appropriate immunizations for encapsulated organisms should be administered prior to performing splenectomy. These immunizations should ideally also be administered prior to rituximab, since anti-CD20 therapy can blunt the immune response to vaccination. While it is optimal to allow sufficient time for a vaccine response, initiation of therapy for ITP should not be delayed while awaiting immune response if therapy is deemed to be required more urgently. Details are presented separately. (See "Elective (diagnostic or therapeutic) splenectomy", section on 'Vaccinations' and "Prevention of infection in patients with impaired splenic function", section on 'Vaccinations'.)

CHOICE OF SECOND-LINE THERAPY

Splenectomy, rituximab, or a TPO-RA — Splenectomy, rituximab, or a thrombopoietin receptor agonist (TPO-RA) are the three principal choices of second-line treatment; they differ in their mechanism of action [4]. All three are effective in raising the platelet count in the majority of individuals, but they differ dramatically in their application, duration, costs, burdens, and adverse effects profiles.

Splenectomy is a one-time permanent surgical procedure that is the most likely of the three principal therapies to result in a durable response; rituximab typically requires four weekly intravenous administrations and may need to be re-administered; TPO-RAs typically require administration for an extended period of time, although responses are sometimes maintained even after stopping treatment. Splenectomy carries operative and other risks. Splenectomy and rituximab both cause immunosuppression (indefinite for splenectomy, potentially prolonged for rituximab).

As a result, the choice of therapy is highly dependent upon patient values and preferences. We discuss all three options with patients and assist them in balancing the risks and benefits of each approach, which are summarized in the table (table 1). A decision tree that incorporates patient values and preferences is illustrated in the algorithm (algorithm 2).

●Splenectomy – Splenectomy may be a good choice for an individual who wishes to have a single potentially curative surgical procedure and who is willing to accept the increased risks of infection and venous thromboembolism. Compared with open splenectomy, laparoscopic splenectomy has a lower surgical mortality and complication rate, shorter hospitalization, and faster recovery. Pre-splenectomy vaccinations are critical. (See 'Decision to pursue splenectomy' below and 'Caveats before proceeding to second-line therapy' above.)

●Rituximab – Rituximab may be a good choice for an individual who wishes to avoid surgery and prefers not to take a medication long term. However, its effect is often short-lived, necessitating redosing or use of another second line agent. (See 'Rituximab evidence for efficacy' below.)

We are trying to avoid rituximab during the coronavirus disease 2019 (COVID-19) pandemic because it can blunt vaccine responses. (See 'Rituximab' below.)

●TPO-RA – A TPO-RA may be a good choice for an individual who is especially concerned about immunosuppression following splenectomy or rituximab and who is less concerned about the need to take a medication for an extended period of time, including the associated costs and burdens. (See 'Decision to use a TPO-RA' below.)

Temporary use of a TPO-RA may be appropriate during the COVID-19 pandemic as a means of avoiding immunosuppressive therapy.

If splenectomy is chosen, it is generally preferable to wait at least one year from the time of diagnosis in case a spontaneous remission occurs. During this time, a TPO-RA may be used temporarily while surgery is being scheduled and preoperative vaccines administered (see 'Immunizations' above). A TPO-RA may also be useful for those who require a temporary increase in platelet count in preparation for splenectomy and do not have an adequate platelet count response to glucocorticoids [5].

The 2019 American Society of Hematology (ASH) guideline on ITP makes weak recommendations for a TPO-RA over rituximab and for rituximab over splenectomy, but the guideline emphasizes the potential usefulness of all three treatments and the importance of patient characteristics (age, ITP history, comorbidities), values, and preferences in the final decision [1].

Some experts prefer other therapies, such as mycophenolate mofetil (MMF) over rituximab, despite having less evidence from randomized trials. (See 'Other therapies' below.)

Relative efficacy and toxicity — Overall efficacy of splenectomy, rituximab, and TPO-RAs is discussed in the sections below. (See 'Decision to pursue splenectomy' below and 'Rituximab evidence for efficacy' below and 'Decision to use a TPO-RA' below.)

These therapies (or any pairwise combinations) have not been directly compared in randomized trials [6]. However, the reported initial response rates with splenectomy are higher than with rituximab (60 to 80 percent versus 55 to 65 percent, respectively) and responses with splenectomy are more durable (often lasting for many years or even indefinitely for splenectomy, versus one to two years for rituximab) [7-13]. There are also significant differences in short-term and long-term risks of splenectomy versus rituximab (table 1) [14,15].

In an unpublished retrospective cohort study comparing TPO-RAs with rituximab in 429 individuals with chronic ITP, survival rates were similar, but toxicities were less with TPO-RAs [16]. Rituximab-related morbidities included infection (pneumonia, sepsis) and thrombotic complications (venous thromboembolism, myocardial infarction). There are significant differences in short-term and long-term risks and burdens of rituximab versus TPO-RAs; namely, rituximab causes immunosuppression and may work only transiently, and TPO-RAs are highly effective but require extended administration in many cases.

Other options — Other therapies may be appropriate if splenectomy, rituximab, and/or a TPO-RA are ineffective or cannot be used. (See 'Other therapies' below.)

SPLENECTOMY — Splenectomy removes the major site of phagocytosis of antibody-coated platelets, as well as lymphocytes that reside in the spleen that might be responsible for producing antiplatelet antibodies. As such, it effectively targets multiple pathophysiologic mechanisms of ITP and has the greatest potential to modify the course of the disease [17]. (See "Splenomegaly and other splenic disorders in adults", section on 'Properties of the normal spleen'.)

Decision to pursue splenectomy — As noted above and summarized in the algorithm (algorithm 2), splenectomy may be a good choice for an individual who wishes to have a single potentially curative surgical procedure and who is willing to accept the small increased risks of infection and venous thromboembolism. Preference for splenectomy may also correlate with the local surgical expertise. (See 'Choice of second-line therapy' above.)

●Efficacy and adverse events – Of all the ITP therapies, splenectomy has the greatest chance of altering the disease course and resulting in a sustained remission [7,18]. However, splenectomy has the greatest associated risk, including operative risks and postoperative risks of immunosuppression and increased rate of thromboembolic complications. (See 'Splenectomy risks' below.)

The high rate of sustained remission and toxicity profile of splenectomy were illustrated in a systematic review of splenectomy for ITP that included data for 47 case series (2623 adults) and over 50 years of observation, published in 2004 [7]. Platelet counts normalized in 66 percent of individuals and increased to >50,000/microL in an additional 22 percent, for a total response rate of 88 percent. Responses persisted throughout the duration of monitoring (up to over 12 years) (figure 1). The complication rates were 13 percent with open splenectomy and 10 percent with laparoscopic splenectomy. Mortality was 1 percent with open splenectomy and 0.2 percent with laparoscopy. (See 'Splenectomy risks' below.)

●Predicting response – The only clinical parameter that predicts a favorable response to splenectomy is patient age; younger patients have a higher response rate, although a specific age cutoff below which splenectomy was most effective could not be determined [7,19-22]. One study reported that the duration of response following splenectomy and the likelihood of response may be reduced in patients older than 65 years [20]. Another study found a correlation between response to intravenous immune globulin (IVIG) and response to splenectomy [23]; however, this may represent better patient selection.

In some centers, a radiolabeled tracer scan (indium, chromium) is used to determine the degree of splenic sequestration of platelets, which is thought to correlate with the likelihood of a good response to splenectomy. A 2020 meta-analysis of studies using this approach in 2966 individuals with ITP demonstrated a correlation between a splenic sequestration pattern and a good response to splenectomy, relative to other patterns (hepatic, mixed, or unknown) [24]. Response rates were 87 percent with a splenic pattern versus 47 percent with a mixed pattern and 26 percent with a hepatic pattern. This type of scanning to predict response to splenectomy is not routinely used in the United States.

●Timing – If splenectomy is pursued, it is generally prudent to wait at least 12 months from diagnosis since spontaneous remissions are possible. (See 'Timing of splenectomy' below.)

Platelet counts typically rise within the first two weeks postoperatively (table 2). Platelet count increases within one to two days can also occur, and delayed responses (up to eight weeks or longer) have been observed [25].

●Accessory spleen – For patients who have persistent thrombocytopenia following splenectomy, we generally proceed to another second-line therapy or, if all second-line options have been exhausted, to a third-line therapy. We do not pursue a diagnosis of an accessory spleen or perform accessory splenectomy in patients with persistent thrombocytopenia following splenectomy. However, this approach has been used in rare patients who have a late recurrence of ITP after splenectomy. Absence of Howell-Jolly bodies (picture 1) on the peripheral blood smear may be a clue to persistent splenic tissue or splenic tissue regrowth. (See "Evaluation of the peripheral blood smear", section on 'Howell-Jolly bodies'.)

Timing of splenectomy — Decisions about when to perform splenectomy can be challenging because of the possibility of a late spontaneous remission. Many clinicians prefer to delay splenectomy with the hope that a spontaneous remission will occur and render the procedure unnecessary [17]. However, data to predict which patients will undergo spontaneous remission (and thus might avoid splenectomy) are lacking, and there is not a clear cutoff time after which spontaneous remissions become less likely.

We individualize the timing of splenectomy based on the patient's clinical features, response to initial therapy, and other considerations (eg, patient preference). In practice, we often wait at least 12 months from diagnosis of ITP to splenectomy, to allow for late spontaneous remissions; this practice is consistent with the 2019 ASH ITP guidelines [1]. Some patients with persistently severe and symptomatic thrombocytopenia despite medical therapy may require splenectomy sooner.

Pre-splenectomy considerations — Additional issues for patients considering splenectomy include the following:

●Immunizations – Patients considering splenectomy should receive immunizations for encapsulated organisms at least two weeks prior to the procedure if possible and if these were not done already. These individuals are often prescribed antibiotics to take in the event of fever. Details of the impact of immunosuppression on response to vaccination, specific vaccination schedules, and other infection prevention are discussed separately. (See "Elective (diagnostic or therapeutic) splenectomy", section on 'Preoperative considerations' and "Prevention of infection in patients with impaired splenic function", section on 'Vaccinations' and "Prevention of infection in patients with impaired splenic function", section on 'Emergency antibiotic supply'.)

●Optimizing the preoperative platelet count – If the platelet count needs to be increased for splenectomy, the patient can be treated with intravenous immune globulin (IVIG), glucocorticoids, or a thrombopoietin receptor agonist (TPO-RA). We prefer to perform splenectomy with a platelet count of ≥50,000/microL; however, many patients with ITP have undergone open or laparoscopic splenectomy safely in the setting of more severe thrombocytopenia, with platelets available for transfusion if urgently needed because of intraoperative bleeding [26]. (See "Initial treatment of immune thrombocytopenia (ITP) in adults", section on 'Surgery or delivery'.)

●Operative technique – Laparoscopic splenectomy has become the standard operative technique [17]. This is based on the lower rates of morbidity and mortality reported in a systematic review that included over 3000 splenectomies in patients with ITP (mortality rate of 1 percent with open and 0.2 percent with laparoscopic procedures) [7]. Practices and surgical outcomes may differ at different institutions. (See "Elective (diagnostic or therapeutic) splenectomy", section on 'Surgical approach'.)

Other general preoperative considerations (eg, cardiac risk assessment, preoperative thyroid function testing) are presented separately. (See "Evaluation of cardiac risk prior to noncardiac surgery" and "Preoperative medical evaluation of the healthy adult patient" and "Nonthyroid surgery in the patient with thyroid disease", section on 'Is preoperative measurement of TSH necessary?'.)

Splenectomy risks — The risks of splenectomy include those of the surgical procedure as well as long-term risks of increased susceptibility to infection and possibly vascular events [27]. In a systematic review that included 47 case series of splenectomy for ITP, complications occurred in 88 of 921 patients who underwent laparoscopic splenectomy (10 percent) and 318 of 2465 who underwent open splenectomy (13 percent) [7]. These and other potential complications are discussed in detail separately. (See "Elective (diagnostic or therapeutic) splenectomy", section on 'Postoperative risks'.)

Risks of splenectomy compared with other second-line therapies are summarized in the table (table 1).

RITUXIMAB — Rituximab is a chimeric monoclonal antibody (mAb) directed against the B cell surface protein CD20. Several biosimilar products have been approved, and we use these interchangeably with rituximab. These antibodies are thought to eliminate B cells via apoptosis, antibody-dependent cytotoxicity, and complement-mediated lysis [28]. Plasma cells responsible for long-term antibody production do not express CD20; this may explain the shorter durations of response relative to response durations seen with splenectomy.

Rituximab is immunosuppressive, and during the coronavirus disease 2019 (COVID-19) pandemic we have tried to avoid using rituximab because it blunts the immune response to vaccinations, including COVID-19 vaccines, probably for at least 6 months.

COVID-19 vaccination should not be withheld for individuals who have recently received rituximab, but antibody responses may need to be assessed and/or repeat vaccination considered.

Patients and clinicians should be informed about the possible reduced vaccine response and the importance of taking other precautions against COVID-19. Additional considerations related to COVID-19 vaccination in individuals receiving rituximab and general precautions are presented separately. Patients and providers are advised to review the latest recommendations with respect to booster vaccine doses in patients who are receiving immunosuppressive therapy. (See "COVID-19: Vaccines", section on 'Immunocompromised individuals' and "COVID-19: Epidemiology, virology, and prevention", section on 'Prevention'.)

Rituximab evidence for efficacy — As noted above and summarized in the flowchart (algorithm 2), rituximab may be a good choice for an individual who wishes to avoid surgery and prefers not to take a medication for an extended period of time; however, the short duration of its effect may mandate retreatment or an additional second-line therapy. (See 'Choice of second-line therapy' above.)

The overall efficacy of single-agent rituximab therapy is in the range of 40 to 60 percent based on data from several large reviews and meta-analyses. The median duration of effect is approximately one year, and some individuals may require retreatment.

●A 2015 meta-analysis of randomized trials that compared rituximab plus standard care versus standard care alone (five trials; 463 patients) found a modest benefit of rituximab in improving platelet counts [29]. A sustained complete platelet count response (platelet count >100,000/microL) at 6 months occurred in 47 percent of patients who received rituximab and 33 percent of controls (relative risk [RR] 1.42, 95% CI 1.13-1.77). A partial response (platelet count above 30,000 to 50,000/microL) occurred in 58 percent of patients who received rituximab compared with 47 percent of controls.

In the trial with the longest follow-up (median, 1.5 years), however, responses with rituximab more closely approached placebo (complete responses in 16 percent of individuals treated with rituximab versus 7 percent treated with placebo; RR 2.21, 95% CI 0.72-6.75) [11]. Avoidance of splenectomy (or avoidance of the criteria for splenectomy [platelet count <20,000/microL or need for continued glucocorticoid therapy]) was seen in 81 percent of the rituximab group and 73 percent of the placebo groups). Three late spontaneous remissions occurred, more than a year after randomization, in patients receiving no therapy.

●A meta-analysis from 2012 that included randomized trials and observational studies in which non-splenectomized individuals were treated with rituximab found an overall response rate of 57 percent and platelet counts >100,000/microL in 41 percent [30].

●A 2007 systematic review of observational studies of (313 adults) found complete responses in 44 percent and platelet count >50,000/microL in 63 percent [9]. Approximately one-half of the individuals had undergone splenectomy, and responses were similar in the prior splenectomy and no splenectomy groups. The median time to response was 5.5 weeks from the first dose of rituximab, and the median response duration was 11 months (interquartile ranges, three to seven weeks and 6 to 18 months, respectively). Toxicities were mild, including grade 1 to 2 infusion reactions and other minor reactions. Ten patients had more serious reactions (eg, infections, allergic reactions, thromboemboli), and there were nine deaths (2.9 percent). Causes of death included infections and bleeding; many of the patients had serious underlying comorbidities.

Although the median duration of response to rituximab (approximately one year) is shorter than splenectomy, some individuals can have long-lasting responses. This was illustrated in a study that evaluated the duration of response in 72 adults who had an initial response to rituximab that lasted for at least one year and 66 children with a response of any duration [8]. After five years of observation, 21 percent of adults and 26 percent of children had a continued response. It is unclear whether these long-lasting responses are attributable to rituximab or whether these cases were destined to remit spontaneously.

Rituximab planning and dosing — Patients should be screened for hepatitis B virus (HBV) infection before starting rituximab because of the risk of HBV reactivation. Rituximab is generally avoided in individuals with active or occult HBV infection; if rituximab must be used, prophylactic anti-HBV treatment against hepatitis B should be considered in consultation with an infectious disease specialist or hepatologist. (See "Hepatitis B virus: Overview of management".)

Treatment with rituximab can suppress vaccine responses for at least six months after administration. Thus, we provide appropriate immunizations prior to starting rituximab therapy [31]. (See "Prevention of infection in patients with impaired splenic function", section on 'Vaccinations'.)

The dose of rituximab typically used in patients with ITP is 375 mg/m² intravenously once a week for four consecutive weeks. We favor this dosing based on more extensive experience with it, although we understand that lower doses may be sufficient for ITP therapy compared with hematologic malignancies. As examples:

●A study that treated 48 patients with rituximab at 100 mg/week for four weeks found similar efficacy to higher doses (response in 60 percent, platelet count normalized in 40 percent) [32,33].

●A study that treated 108 patients with two fixed doses of rituximab (1000 mg on days 1 and 15) showed a response rate of 44 percent [34].

●A study that treated 248 patients with either 375 mg/m² once a week for four weeks or 1000 mg on days 1 and 15 (choice of regimen based on clinician preference) showed similar efficacy of the two regimens, with initial responses in 152 patients (62 percent and 61 percent) and lasting responses at two years in 96 patients (39 percent and 40 percent) [35].

Responses to rituximab can occur within a week but may take up to two months (table 2). Rituximab may be administered in combination with other treatments because its effects on platelet count can be delayed.

Rituximab toxicities — Adverse effects of rituximab include infusion reactions and prolonged immunosuppression, which can result in reactivation of HBV infection [35]. Progressive multifocal leukoencephalopathy (PML) has been reported following rituximab for ITP; many of the affected individuals had been pretreated with other immunosuppressive agents in addition to rituximab [36]. Boxed warnings for infusion reactions, hepatitis B reactivation, mucocutaneous reactions, and PML are included in the prescribing information. These toxicities are discussed in more detail separately. (See "Rituximab: Principles of use and adverse effects in rheumatoid arthritis", section on 'Adverse effects' and "Hepatitis B virus reactivation associated with immunosuppressive therapy".)

As noted above, rituximab has been reported to interfere with response to immunizations, which is of potential concern in patients who may subsequently undergo splenectomy; we provide immunizations prior to administering rituximab. (See 'Immunizations' above.)

The mechanisms by which rituximab may cause immunosuppression are discussed separately. (See "Secondary immunodeficiency induced by biologic therapies", section on 'Rituximab'.)

TPO RECEPTOR AGONISTS — Thrombopoietin receptor agonists (TPO-RAs), also called TPO mimetics, include small molecule peptide and non-peptide agents. These act by stimulating the production of megakaryocytes and ultimately platelets in the bone marrow by binding to and activating the TPO receptor. (See "Biology and physiology of thrombopoietin".)

Decision to use a TPO-RA — As noted above and summarized in the flowchart (algorithm 2), a TPO-RA may be a good choice for an individual who wishes to avoid surgery and the immunosuppressive effects of splenectomy or rituximab and who is less concerned about the need to take a medication for an extended period of time. (See 'Choice of second-line therapy' above.)

Approximately 80 percent of individuals with ITP will have a significant platelet count increase in response to a TPO-RA [37,38]. Several reports have documented sustained responses in some individuals with ITP after discontinuation of a TPO-RA, with response rates up to 30 to 50 percent, although observation periods have been relatively short (<1 year) [17,39,40]. In most individuals, these agents do not induce remission, and prolonged maintenance therapy is usually required. If treatment is discontinued, platelet counts generally return to baseline levels or even below baseline ("rebound thrombocytopenia").

Choice of TPO-RA — Available TPO-RAs for ITP include romiplostim, eltrombopag, and avatrombopag. They are all effective, but their relative efficacy in ITP has not been evaluated in randomized trials. The 2019 American Society of Hematology (ASH) guideline does not express a preference for one TPO-RA and suggests a choice between romiplostim or eltrombopag based on patient preferences for route of administration [1]. We individualize the choice based on availability, cost, patient comorbidities, and patient preference. Some individuals have a strong preference for one administration route (oral or subcutaneous) over the other.

●Romiplostim (Nplate) is administered as a once-weekly subcutaneous injection. It is a recombinant protein that contains a peptide with four binding sites for the TPO receptor linked to an IgG1-Fc component, termed a "peptibody" [41-43]. (See 'TPO-RAs dosing and monitoring' below.)

●Eltrombopag (Promacta, Revolade) and avatrombopag (Doptelet; approved for ITP in 2018) are given as a once-daily pill. They are non-peptide small molecules that activate the TPO receptor.

•Eltrombopag should be dose reduced or avoided in patients with liver disease or increased liver enzymes. Dose reductions are recommended for patients of East Asian descent. Prescribing information includes a Boxed Warning regarding risks of hepatic decompensation in individuals with chronic hepatitis taking ribavirin or interferon and risks of hepatotoxicity. It is taken without food or with food low in calcium. (See 'TPO-RAs dosing and monitoring' below.)

•Avatrombopag can be taken with food. (See 'TPO-RAs dosing and monitoring' below.)

Recombinant human TPO has been used in China [44,45]. Recombinant pegylated TPO products such as recombinant PEGylated megakaryocyte growth and development factor (PEG-rHuMGDF) are not used in patients with ITP, following reports of the development of anti-TPO antibodies in response to one of the products that caused severe thrombocytopenia. (See "Clinical applications of thrombopoietic growth factors", section on 'Recombinant thrombopoietins' and "Clinical applications of thrombopoietic growth factors", section on 'Antibody formation'.)

TPO-RAs dosing and monitoring — TPO-RAs are titrated to keep the platelet count in a safe range (typically, ≥50,000/microL) without causing thrombocytosis. Dosing starts with a low initial dose that is adjusted as needed, with close monitoring of the platelet count until a stable dose is reached.

●Dosing

•Romiplostim – Subcutaneous injection once-weekly. Typical starting dose, 2 to 3 mcg/kg. Most patients in the clinical trials required 3 to 8 mcg/kg (range from 1 to 10 mcg/kg) [46,47].

Vials contain either 250 mcg or 500 mcg of the drug, which would equal approximately 3 mcg/kg or 6 mcg/kg for an 80 kg patient. [46].

•Eltrombopag – Oral pill once daily. Absorption is reduced by polyvalent cations; eltrombopag should be taken without a meal or with a meal low in calcium (≤50 mg) and at least two hours before and four hours after calcium-containing foods, medications such as antacids, and supplements containing polyvalent cations such as calcium, iron, aluminum, magnesium, selenium, or zinc.

Initial dose, 50 mg daily, or 25 mg once daily in individuals of East-Asian ancestry (Chinese, Japanese, Korean, Taiwanese) and individuals with moderate or severe liver insufficiency [48].

Eltrombopag inhibits the transporter, OATP1B1, which is responsible for transporting certain drugs such as rosuvastatin. Patients should be monitored closely for toxicities of OATP1B1 substrates and/or they may decrease the dose of these drugs.

•Avatrombopag – Oral pill once daily, taken with food. There are no interactions with dairy products or other medications and there is less risk of hepatotoxicity compared with eltrombopag.

●Platelet count monitoring – For all of these agents, platelet counts increase in approximately 7 to 14 days (table 2) [49].

To avoid risks of thrombocytosis, we generally check the platelet count:

•Approximately one week after starting therapy

•Approximately one week after dose changes

•Monthly after a stable dose is reached

•If there is a concern about thrombocytopenia (eg, petechiae, mucosal bleeding, intercurrent illness)

●Other monitoring – As noted, the following additional laboratory tests are appropriate:

•Liver enzymes – For eltrombopag, liver enzymes should be measured prior to drug initiation, every two weeks during dose adjustments, and monthly thereafter.

TPO-RAs efficacy and adverse events — As noted above, TPO-RAs have not been directly compared in individuals with ITP. (See 'Choice of TPO-RA' above.)

The efficacy and toxicities of individual TPO-RAs have been evaluated in various clinical trials such as the following:

●Romiplostim – The efficacy of romiplostim in raising platelet counts is approximately 80 percent, with higher response rates when transient responses are included [50-53]. The following studies illustrate the range of findings:

•A pooled analysis of data from 14 trials involving 1059 patients treated with romiplostim for up to five and a half years found no difference in the rates of thrombosis, hematologic malignancy/dysplasia, or non-hematologic malignancies over that seen in controls [54]. Bone marrow reticulin was seen in 17 of 921 patients receiving romiplostim and 1 of 65 patients receiving placebo (1.8 versus 1.5 percent); an additional romiplostim-treated patient had increased collagen. Other series have reported rates of adverse events in the range of 13 percent, with thrombotic events in the range of 5 to 7 percent [46,54,55].

•An open-label randomized trial from 2010 in 234 patients with ITP found a higher rate of platelet count responses with romiplostim (platelet count >50,000/microL: 71 to 92 percent with romiplostim versus 26 to 51 percent with standard care); the proportion undergoing splenectomy was also lower (9 percent with romiplostim versus 36 percent with standard care) [47]. Serious adverse events occurred in 23 percent of romiplostim-treated patients versus 37 percent of controls.

•A randomized trial from 2008 in 125 adults with ITP found responses to romiplostim in 38 percent of individuals with prior splenectomy and 61 percent without prior splenectomy, versus 0 to 5 percent with placebo [51]. Additional romiplostim-treated patients had transient responses. Response to romiplostim was accompanied by decreased bleeding and an improvement in health-related quality of life as measured by the ITP-Patient Assessment Questionnaire [56].

•A series of 292 adults treated with romiplostim for up to five years found responses in 90 percent [46]. Responses were inversely correlated with the duration of ITP. Overall, 67 percent of patients were able to avoid use of other ITP medications such as glucocorticoids. Headache was the most common adverse event.

●Eltrombopag – The efficacy of eltrombopag in raising platelet counts is in the range of approximately 80 percent, with higher response rates when transient responses are included [57-59].

•A 2011 trial (RAISE) randomly assigned 197 adults with chronic ITP and platelet count <30,000/microL to eltrombopag or placebo and found platelet count responses in 79 percent of eltrombopag-treated patients (versus 28 percent with placebo) [58]. Durable responses occurred in 51 percent of individuals who had previously undergone splenectomy and 66 percent who had not undergone splenectomy (versus 10 percent of controls). Clinically significant bleeding was reduced (33 percent, versus 53 percent with placebo).

•The 2013 EXTEND study followed 299 patients treated with eltrombopag for up to three years. Responses were seen in 80 and 88 percent of patients with and without splenectomy, and responses were maintained (with continued administration of eltrombopag) for approximately 70 percent of the study duration [59]. A follow-up of EXTEND published in 2017 reported that the median duration of eltrombopag was approximately 2.4 years, and 259 of 302 patients (86 percent) had a platelet count >50,000/microL at least once [60]. Adverse events leading to withdrawal (hepatobiliary changes, cataracts, thrombosis, headache, myelofibrosis) occurred in 41 (14 percent), but rates of hepatobiliary and thromboembolic events did not increase with treatment duration beyond one year.

•In a 2011 trial that included 197 patients, 17 of 135 (13 percent) discontinued therapy because of adverse events, the most common of which were liver enzyme abnormalities (four patients) and thromboembolic events (two patients) [58]. There were no significant differences in the development of cataracts or malignancies in the eltrombopag and placebo groups. Additional studies have not demonstrated other safety concerns [46,59,61,62].

●Avatrombopag – There is less clinical experience with avatrombopag, which was approved by the US Food and Drug Administration in 2018.

•In a 2014 trial in 64 adults with chronic ITP (relapsed or unresponsive to first-line therapy), avatrombopag showed a dose-dependent increase in platelet counts (percentage of individuals with a response: 0 responses with placebo, 13 percent with 2.5 mg daily, 53 percent with 5 mg daily, 50 percent with 10 mg daily, and 80 percent with 20 mg daily) [63]. Four individuals had thromboembolic events; most had other risk factors for thromboembolic disease.

•In a trial from 2018 in 49 adults with chronic ITP and platelet counts <30,000/microL, avatrombopag at 20 mg daily resulted in a greater duration of sustained platelet count ≥50,000/microL than placebo (median, 12.4 versus 0 weeks) [64]. There was a very high dropout rate, attributed to lack of a platelet count response, especially in the placebo group, in which only one of 17 completed the full trial. Rates of bleeding and use of rescue medications were similar between groups when adjusted for the difference in length of participation. There were four thromboembolic events; three of these individuals were reported to have risk factors for thromboembolism.

•Preclinical studies in animals suggested that excessively high doses of avatrombopag might cause carcinoid tumors related to changes in gastrin levels; however, in clinical trials, gastrin levels were normal and no clinical issues were seen [65].

Most studies have demonstrated minimal serious adverse events even after administration of these agents for extended periods of time [46]. However, concerns related to the following potential toxicities have been raised:

●Thrombosis – All TPO-RAs have been associated with a small increased risk of thrombosis in some populations. However, results from a 2015 meta-analysis suggest that the risk of thrombosis was not increased by romiplostim or eltrombopag in patients with ITP [66]. Smaller studies with avatrombopag suggest that thrombosis is not a risk with short-term treatment, but there are insufficient data to determine the actual risk [67].

●Bone marrow reticulin – Initial concerns were raised regarding the possibility of increasing bone marrow reticulin formation with romiplostim and eltrombopag; however, this was not demonstrated in follow-up studies [54]. We do not monitor bone marrow fibrosis in individuals receiving a TPO-RA, as there is no evidence of progression to myelofibrosis and no evidence that serial bone marrow testing leads to improved outcomes. In a romiplostim study using bone marrow surveillance of 131 biopsies, nine (7 percent) demonstrated increases of reticulin and collagen of ≥2 grades on the modified Bauermeister scale after one to three years of treatment [68]. Three of the nine patients had repeat biopsies after stopping the drug that showed resolution of the abnormalities. In an eltrombopag study with 5.5 years of follow-up, moderate reticulin fibrosis of the bone marrow was seen in 2 of 117 patients (1.7 percent) [69].

●Antibodies – These agents have no sequence homology to thrombopoietin and are not expected to generate inhibitory antibodies.

These issues and potential toxicities in patients with other underlying conditions (eg, theoretical risk of hematologic malignancy in patients with myelodysplasia) are discussed in more detail separately. (See "Clinical applications of thrombopoietic growth factors", section on 'Side effects and risks'.)

OTHER THERAPIES — For patients for whom the therapies discussed above are ineffective, not well tolerated, or unavailable, management may be more complex, and a hematologist with experience in managing patients with ITP should be involved in the treatment decisions.

A number of options are available, as discussed in the following sections. Choices among these therapies vary according to clinician and patient preference, and it may take time to find the best agent for each individual patient. Among the third-line therapies, we generally prefer fostamatinib because it is supported by higher-quality evidence than the other options.

Fostamatinib — Fostamatinib (also called fostamatinib disodium or fostamatinib disodium hexahydrate) was approved by the US Food and Drug Administration in mid-2018 for the treatment of thrombocytopenia in adults with chronic ITP who have had an insufficient response to a previous treatment [70].

Fostamatinib is a small molecule prodrug of a tyrosine kinase inhibitor that inhibits Syk (spleen tyrosine kinase). The major metabolite of fostamatinib, R406, is active in Syk inhibition. R406 is thought to increase the platelet count in patients with ITP by reducing phagocytosis and destruction of autoantibody-coated platelets by macrophages, via inhibition of signal transduction through Fc-activating receptors and the B-cell receptor [71,72].

●Dosing – The initial dose for fostamatinib is 100 mg orally twice daily, which can be increased to 150 mg twice daily if the platelet count has not increased to at least 50,000/microL after one month [73]. The complete blood count (CBC) including platelet count should be monitored monthly until stable; liver function tests are monitored monthly; and blood pressure is monitored every two weeks until a stable dose and monthly thereafter. Dose escalations or reductions, drug discontinuation, or the addition of other therapies (eg, antihypertensives, antiemetics) may be required. Like TPO-RAs, this is expected to be a maintenance treatment that must be continuously administered to maintain a safe platelet count.

●Side effects – Side effects of fostamatinib include diarrhea, nausea, hypertension, neutropenia, and rash [74,75].

●Supporting evidence – Evidence for fostamatinib is more limited than for splenectomy, rituximab, or TPO-RAs, but high-quality evidence for efficacy and safety is emerging, making it a reasonable choice among subsequent treatment options. Its role may change (it may become a second-line choice) as more data become available.

•In two randomized trials (FIT-1 and FIT-2) from 2018, which together randomly assigned 150 individuals with chronic ITP (median duration, 8.5 years; median platelet count, 16,000/microL) to receive fostamatinib or placebo, fostamatinib was associated with a greater rate of stable platelet count >50,000/microL (18 versus 2 percent; p = 0.0003) [74]. Forty-four individuals assigned to placebo underwent crossover to fostamatinib (FIT-3 study), and 10 of these (23 percent) also had a platelet count response. The median time to response was approximately two weeks. Severe and serious bleeding events were reduced with fostamatinib (5 versus 16 percent). The most common adverse events were diarrhea, hypertension, nausea, dizziness, and an increase in alanine aminotransferase (ALT); less common adverse events included neutropenia and rash.

•In an open-label extension study, participants in FIT-1 and FIT-2 were followed over a longer duration, stable responses with a median duration of >28 months occurred in 18 percent [75].

•In the initial 2009 dose-finding study, 16 adults with chronic ITP refractory to at least two other treatments with platelet counts <30,000/microL were treated with fostamatinib at increasing doses starting at 75 mg twice daily and increasing up to 150 mg twice daily [76]. Responses, defined as a platelet count increase of at least 20,000/microL, were seen in 12 individuals (75 percent; median peak platelet count 100,000/microL), and responses were sustained in eight (50 percent), some of whom were able to reduce or discontinue other ITP therapies such as intravenous immune globulin (IVIG) or glucocorticoids. Bleeding was not reported. Adverse events were mostly gastrointestinal (eg, diarrhea with increased urgency, nausea, vomiting, increased hepatic transaminases). Three patients discontinued therapy due to adverse effects.

While the efficacy of fostamatinib was modest, many of the patients who had a response had previously tried many other treatments that were not effective, including splenectomy and thrombopoietin receptor agonists (TPO-RAs). In the FIT trials, patients had tried at least one prior therapy, and most had tried multiple (median, 3; range, up to 13) [74].

Immunosuppressive agents — Immunosuppressive agents other than glucocorticoids at the standard dose and schedule have been used to treat ITP, with case reports or small, uncontrolled series describing successful use [19]. However, these case series often included patients with mild short-term thrombocytopenia (ie, possibly not ITP) or had only a minority of patients with ITP unresponsive to splenectomy, rituximab, or a TPO-RA [77].

●Intermittent glucocorticoids – Although glucocorticoids are highly effective as initial therapy, we generally avoid long-term continuous and intermittent glucocorticoids due to toxicities with extended use and availability of other good options. An exception is the uncommon patient who is able to maintain a safe platelet count on very low-dose glucocorticoids (prednisone at a dose of ≤5 mg/day). (See "Initial treatment of immune thrombocytopenia (ITP) in adults", section on 'Glucocorticoids and IVIG'.)

We sometimes use low-dose, intermittent glucocorticoids in patients who have a platelet count response. As an example, some patients maintain a safe platelet count on oral prednisone at doses as low as 5 or 10 mg every other day. However, even these low doses can accelerate the development of osteoporosis and other glucocorticoid toxicities [78-80]. (See "Major adverse effects of systemic glucocorticoids".)

●MMF – A study involving 46 individuals with severe ITP reported disease responses to mycophenolate mofetil (MMF) in 24 (52 percent); 15 had platelet counts over 100,000/microL [81]. The standard dose was 1 g daily. MMF has also been used in combination regimens. (See 'Combinations' below and "Initial treatment of immune thrombocytopenia (ITP) in adults", section on 'Other therapies and multiagent combinations'.)

●Azathioprine – A series of 53 individuals with chronic ITP that persisted despite other therapies (splenectomy in 40) reported success with azathioprine at an initial oral dose of 150 mg/day; in some cases, several months of therapy was needed before a response was seen [82]. Thiopurine methyltransferase (TPMT) testing should be performed in all patients initiating azathioprine.

●Cyclosporine – Therapies used by other clinicians in extreme circumstances with reports of success have included cyclosporine, vincristine, and cyclophosphamide, alone or in combination, and hematopoietic stem cell transplant [83-88].

●IVIG – We consider IVIG to be rescue treatment in ITP rather than routine, ongoing therapy. IVIG is important for patients with clinically important bleeding and for patients in whom a more rapid response is required [19]. (See "Initial treatment of immune thrombocytopenia (ITP) in adults", section on 'IVIG dosing and administration'.)

Danazol and dapsone

●Danazol – Danazol is an attenuated androgen that may have efficacy in ITP. Danazol is generally well tolerated in males, but hirsutism can be a troubling side effect for females. In uncontrolled case series, platelet count responses to initial oral doses of 600 mg/day have been reported in up to 72 percent of patients [89-91]. Its efficacy in patients with chronic refractory ITP is unknown and may be limited [19,92].

●Dapsone – A series of 122 individuals treated in a referral center with dapsone reported responses in 81 (66 percent) [93]. The starting dose was 100 mg daily; higher doses did not improve the response rate. The cohort mostly consisted of individuals with chronic ITP, although some were treated with dapsone as their initial ITP therapy. Complete responses were seen in 29 (24 percent), with sustained responses in approximately one-half. Mild elevations in methemoglobin were common, with symptomatic methemoglobinemia in five individuals. (See "Methemoglobinemia", section on 'Dapsone'.)

Combinations — In some patients who require therapy because of severe and symptomatic thrombocytopenia unresponsive to single agents, combinations of agents may be effective. The rationale for giving a combination of therapies include the need for more rapid effects of the faster-acting agents until the more durable effects of longer-acting agents take effect and the possibility of synergism between different agents [94].

Combinations that have been tested include:

●Various combinations with a TPO-RA, especially with an agent that inhibits platelet clearance, such as fostamatinib, rituximab, or a glucocorticoid.

●Triple therapy given over 4 weeks (TT4) using high-dose oral dexamethasone (40 mg daily on days 1 to 4), oral cyclosporine (2.5 to 3 mg/kg daily on days 1 to 28) and low-dose rituximab (100 mg on days 7, 14, 21, and 28) [95].

●Azathioprine, MMF, and cyclosporine [96].

●Glucocorticoids, IVIG, vincristine, and/or anti-D was administered to 35 patients whose disease did not respond to single-agent glucocorticoids, IVIG, or splenectomy, with or without azathioprine and danazol [97].

●Glucocorticoids, cyclophosphamide, and vincristine, procarbazine, or etoposide [98].

●All-trans retinoic acid (ATRA) with low dose rituximab [99].

For the most part, adverse events in these studies have been considered tolerable, but they are generally greater than seen in patients treated with single agents [95-98,100,101].

INVESTIGATIONAL THERAPIES

Rilzabrutinib (investigational BTK inhibitor) — Rilzabrutinib is an oral Bruton tyrosine kinase (BTK) inhibitor under investigation for several immune-mediated disorders.

In a 2022 open label dose-finding study, 60 patients with ITP and platelet count <30,000/microL not responsive to other therapies were treated with one of several rilzabrutinib dosing regimens, with the option for dose escalation [102]. The median duration of ITP was six years, and patients had received a median of four prior therapies. Response, defined as platelet count >50,000/microL and at least 20,000/microL over baseline on at least two occasions, occurred in 24 patients (40 percent). The median time to platelet count response was approximately 12 days. One-half of the participants had a treatment-related adverse event; all were mild (nausea, diarrhea, fatigue).

Increased specificity is thought to reduce the risk of off-target effects such as atrial fibrillation. Other BTK inhibitors can cause bleeding by inhibiting platelet aggregation, but rilzabrutinib does not appear to have this effect.

Efgartigimod (investigational FcRn inhibitor) — Efgartigimod reduces the levels of anti-platelet autoantibodies that may contribute to the pathophysiology of ITP; it does so by acting as a decoy for the neonatal Fc receptor (FcRn), in turn preventing FcRn from stabilizing circulating IgG [103].

In a 2023 randomized trial involving 131 adults with chronic or persistent ITP and platelet count <30,000/microL, efgartigimod 10 mg/kg intravenously was more effective than placebo in raising the platelet count above 50,000/microL (22 versus 5 percent) [104]. The patients had longstanding ITP, with a mean of 10.6 years since diagnosis; two-thirds had received at least three prior ITP therapies, and slightly over one-third had a splenectomy. Therapy was well-tolerated, but regular intravenous infusions (once weekly or every other week) are required.

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Immune thrombocytopenia (ITP) and other platelet disorders".)

SUMMARY AND RECOMMENDATIONS

●Threshold for therapy – The goal of therapy is to provide a safe platelet count to prevent clinically important bleeding. We often initiate second-line therapy at a lower platelet count (<20,000/microL) than first-line therapy (<30,000/microL) (algorithm 1). Therapy is more important for patients with a higher risk of bleeding (comorbidities such as chronic kidney disease, concomitant antithrombotic agents, athletic activities). (See 'Initial considerations' above and "Initial treatment of immune thrombocytopenia (ITP) in adults".)

●Before starting – Confirm the diagnosis, evaluate for treatable causes of secondary immune thrombocytopenia (ITP), and provide appropriate vaccinations prior to immunosuppressive therapy. (See 'Initial considerations' above.)

●Choice of second-line therapy – The principle options are splenectomy, rituximab, or a thrombopoietin receptor agonist (TPO-RA). These differ in mechanism of action, efficacy, administration, and adverse effects; the choice is individualized. We discuss risks and benefits of all options (table 1) and assist the patient in arriving at a decision consistent with their values and preferences (algorithm 2). We avoid intravenous immunoglobulin (IVIG), except in special circumstances, given its costs, complexity of administration, potential side effects, and transient effect.

We are also trying to avoid immunosuppressive therapy when possible during the coronavirus disease 2019 (COVID-19) pandemic due to possible blunting of the immune response to COVID-19 vaccination. Receipt of immunosuppressive therapy has implications for COVID-19 vaccination. (See 'COVID-19 vaccination' above and "COVID-19: Vaccines", section on 'Immunocompromised individuals'.)

•Splenectomy – May be a good choice for an individual who wishes to have a potentially curative procedure (response rate, 80 to 90 percent) and avoid a regular medication, and who accepts risks of immunosuppression and venous thromboembolism (VTE). Predictors of efficacy include younger age and splenic sequestration on radiolabeled liver-spleen scan (not used routinely in the United States). The laparoscopic approach carries lower surgical morbidity and mortality. (See 'Splenectomy' above.)

-We often wait at least 12 months to allow time for late spontaneous remissions to occur.

-We provide preoperative immunizations.

-We suggest a laparoscopic rather than an open procedure (Grade 2C).

•Rituximab – May be a good choice for an individual who wishes to avoid surgery, prefers not to take a medication for an extended duration of time, and is willing to accept risks of immunosuppression and possible need for retreatment. Efficacy is 40 to 60 percent; median duration is approximately one year. Over the long-term follow-up, its efficacy is similar to placebo, and some individuals may require retreatment. (See 'Rituximab' above.)

-We provide immunizations and test hepatitis B virus (HBV) serologies prior to administration.

-The optimal dose is unknown; many clinicians use 375 mg/m² intravenously once a week for four consecutive weeks.

•TPO-RA – May be a good choice for an individual who wishes to avoid surgery and immunosuppression and is less concerned about the costs and burdens of taking a daily or weekly medication. Some individuals have a strong preference for the oral or subcutaneous route; of the oral agents, experience is greater with eltrombopag than avatrombopag. Adverse effects include VTE and bone marrow fibrosis (reversible and generally not clinically significant). Close platelet count monitoring is used to avoid thrombocytosis. Eltrombopag requires liver enzyme testing. These agents are expensive. (See 'TPO receptor agonists' above.)

-Romiplostim – Weekly subcutaneous injection.

-Eltrombopag – Oral daily pill taken without polyvalent cations.

-Avatrombopag – Oral daily pill taken with food.

●Third-line therapies – Options include fostamatinib, immunosuppressive agents, danazol, dapsone, and combinations. Choices are individualized; it may take time to find the best approach for each patient. (See 'Other therapies' above.)

●Children and pregnancy – (See "Immune thrombocytopenia (ITP) in children: Initial management" and "Immune thrombocytopenia (ITP) in children: Management of chronic disease" and "Thrombocytopenia in pregnancy", section on 'Immune thrombocytopenia (ITP)'.)

●Diagnosis and initial therapy – (See "Immune thrombocytopenia (ITP) in adults: Clinical manifestations and diagnosis" and "Initial treatment of immune thrombocytopenia (ITP) in adults".)

ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges James N George, MD, who contributed to many earlier versions of this topic review.