FormalPara Key Points

Thrombopoietin receptor agonists (TPO-RAs) offer different clinical opportunities based on their pharmacokinetic, pharmacodynamic, and mechanistic characteristics.

In addition to the safety and efficacy of each agent, pharmacists counseling patients on TPO-RA use should consider patient preference for route of administration, the need for stable platelet control and impact of dietary restrictions.

An in-depth knowledge of the TPO-RA class will empower pharmacists and other healthcare professionals to participate in appropriate decision making with individual patients.

1 Introduction

Understanding of thrombopoiesis, or the formation of the platelet component of blood, has improved over time, with the recognition of thrombopoietin (TPO) as the primary regulator of platelet production, renewal, and expansion [1]. Although the existence of TPO was first described in the literature in 1958, it was not until 1994 that basic research uncovered the gene encoding the receptor, c-Mpl, which led to the discovery of its endogenous ligand, TPO [1,2,3]. These discoveries, paired with the knowledge of impaired thrombopoiesis in patients with immune thrombocytopenia (ITP), led to research on TPO receptor agonists (TPO-RAs) [4, 5] and subsequent clinical trials leading to their approval [6,7,8,9,10,11].

This review first provides an in-depth look at available TPO-RA options for second-line treatment of ITP in the adult and pediatric populations and addresses issues that are relevant from a pharmacist’s perspective, including their therapeutic and safety profile, and key considerations for engaging patients in the decision-making process. This review also discusses the use of TPO-RAs in severe aplastic anemia, adults with chronic liver disease who are scheduled to undergo a procedure, and off-label uses such as use following hematopoietic stem-cell transplant, in myelodysplastic syndromes, and in chemotherapy-induced thrombocytopenia.

2 Thrombopoietin Receptor Agonists (TPO-RAs) in Immune Thrombocytopenia (ITP)

Pathologically, ITP is classically characterized by an increase in platelet destruction and decrease in platelet production (due to inhibition of megakaryocyte function) [12, 13]. ITP is a diagnosis of exclusion, relying on clinical and laboratory indicators including a platelet count <100 × 109/L. A definitive diagnosis of ITP also requires exclusion of other causes of thrombocytopenia [14]. Acute ITP, or newly diagnosed ITP, may be characterized by disease < 3 months from diagnosis; persistent ITP refers to disease that is 3 months to 1 year from diagnosis; and chronic ITP refers to thrombocytopenia persisting for 12 or more months after diagnosis [15, 16]. The International Working Group (IWG) classifies ITP as refractory when primary ITP has been confirmed by exclusion of other potential secondary causes of thrombocytopenia, patients are unresponsive to splenectomy or experience relapse following splenectomy, and patients require rescue therapy for clinically significant bleeding [16].

Although not all patients with ITP require second-line therapy, current US and international guidelines recommend use of TPO-RAs as one option among several others (e.g., rituximab, splenectomy) for second-line treatment of ITP following first-line treatment with intravenous immunoglobulin G (IV IgG), corticosteroids, or anti-D immune globulin [4, 5]. In evaluating TPO-RAs, pharmacists should take into account expected response rates, onset of action, patient-specific dosing, cost, considerations for combination therapy in treatment-refractory patients, relevant safety characteristics, and, on occasion, evidence in support of switching agents when current therapy fails to meet goals [17].

2.1 Thrombopoietin and Platelet Life Cycle

The TPO receptor (also known as c-Mpl or CD110) is a 635-amino-acid protein encoded by the c-Mpl gene. The TPO receptor has three functional domains including an extracellular cytokine binding region, a transmembrane domain, and a cytoplasmic domain that binds to signaling molecules including Janus kinases (JAKs) and signal transducer and activator of transcription proteins (STATs) [18]. The TPO receptor ligand, thrombopoietin, is produced primarily in the liver but also in the kidneys and bone marrow. It binds to TPO receptors on megakaryocytes, hematopoietic stem cells, and platelets themselves [2, 19, 20]. TPO stimulates proliferation and differentiation of megakaryocytes, resulting in increased platelet production [18, 21]. Regulation of endogenously produced thrombopoietin levels relies on a feedback loop involving the Ashwell-Morrell receptor expressed on hepatocytes. As a result of reduced steric hindrance, these receptors gain the ability to recognize aged platelets that have lost sialic acid residues. This resultant clearance of desialylated platelets from circulation by hepatocytes triggers hepatic production of thrombopoietin [22].

The pathophysiology of ITP is thought to primarily disrupt the platelet life cycle through generation of autoantibodies and platelet-reactive CD8+ T cells, which act to both reduce platelet production and increase clearance of pre-existing platelets from circulation. Macrophages and dendritic cells may also be involved in sensitizing T cells to platelet fragments [19, 21, 23,24,25]. Although romiplostim competes for the same receptor site as endogenous thrombopoietin, eltrombopag, avatrombopag, and lusutrombopag bind to the transmembrane region of the TPO receptor to activate the TPO receptor noncompetitively [26,27,28,29]. Pharmacologic characteristics of TPO-RAs are summarized in Table 1 [6, 8, 17, 27,28,29,30,31,32,33,34].

Table 1 Comparison of TPO-RAs [6, 8, 17, 27,28,29,30,31,32,33,34]
Full size table

2.2 Potency Comparisons of TPO-RAs in ITP

In ITP, the relative potency of TPO-RAs differs based on patient- and treatment-related factors. Specifically, in patients with ITP, lower endogenous TPO levels (≤ 100 pg/mL as assessed by an enzyme-linked immunosorbent assay) predict for an increased probability of response to eltrombopag and romiplostim, as compared with higher TPO levels (> 200 pg/mL) [17]. This is consistent with an analysis of phase III data (ClinicalTrials.gov identifier NCT01438840) with avatrombopag showing a relationship between efficacy and baseline TPO levels. In this analysis, there was a numerical trend in baseline platelet count and response (i.e., platelet level ≥ 50 × 109/L that is also 20 × 109/L higher than baseline) decreasing from 71% in 18 patients with low TPO levels (< 115 pg/mL), 45% in eight patients with normal TPO levels (115–228 pg/mL), and 15% in five patients with high TPO levels (> 228 pg/mL) [35]. Peak platelet counts with maximal doses of romiplostim were 8–10 times higher than those achieved with eltrombopag, and peak platelet counts with maximal doses of avatrombopag were 3–5 times higher than those achieved with eltrombopag [17].

Although peak platelet counts may be higher in patients receiving romiplostim versus the orally administered TPO-RAs avatrombopag and eltrombopag, exaggerated pharmacologic effects with romiplostim may lead to wide swings in platelet counts, resulting in difficulties in dosage adjustment. These difficulties are compounded by the fact that romiplostim is only approved for use in health care settings in the United States, whereas approval extends to home administration in the European Union [31, 36]. Gilreath et al. presented two cases of patients with chronic ITP in whom stable platelet counts were achieved using a modified dosing regimen of romiplostim. In each case, following romiplostim drug label dosing recommendations led to unexpected supratherapeutic platelet counts as well as subtherapeutic platelet counts that required rescue therapy (IV IgG) for severe thrombocytopenia. In patients with supratherapeutic platelet counts, rather than discontinuing treatment, the authors decreased the dose by 10–20% rather than a set μg/kg decrement. Authors based this on the trend established by prior platelet count data over the preceding weeks (in one case, using a dose of 6.4 µg/kg weekly) [37]. Current recommendations and prescribing information for romiplostim do not take into account trends in platelet count reduction and currently recommend interrupting romiplostim dosing for patients with platelet counts > 400 × 109/L [31]. Pharmacists aware of this pharmacokinetic/pharmacodynamic disconnect may be able to help avoid rebound thrombocytopenia in some patients.

2.3 Efficacy of TPO-RAs in ITP in Adults

2.3.1 Methods

Available TPO-RAs approved for use in ITP include avatrombopag, eltrombopag, and romiplostim. In each of the phase III trials establishing treatment safety and efficacy, the double-blind core treatment period was 6 months in duration [6,7,8]. In all three trials, efficacy evaluations included rates of the median cumulative weeks of response (cumulative weeks with platelet counts ≥ 50 × 109/L), early platelet response (platelet count ≥ 50 × 109/L at day 8), and durable response (platelet counts ≥ 50 × 109/L in ≥ 6 of the final 8 weeks of treatment during the core phase). Although there were some similarities in trial design, notable differences between end point definitions across trials did exist [6,7,8]. For example, patients who had received prior rescue therapy (i.e., use of any additional ITP therapy) were disqualified from being counted as weekly responders during the avatrombopag study [6], but could be counted as weekly responders in the eltrombopag trial regardless of prior rescue therapy use [7], and could be counted as weekly responders in patients who used rescue therapy > 8 weeks prior to response in the romiplostim trial [8].

2.3.2 Avatrombopag

In the phase III study of avatrombopag, 49 patients were randomized in a 2:1 ratio to avatrombopag 20 mg daily or placebo. For the primary outcome of the avatrombopag pivotal trial, the median cumulative weeks of response without rescue therapy was 12.4 weeks with avatrombopag versus 0 weeks with placebo (p < 0.0001). Early platelet response, which was a predefined secondary end point defined as a platelet count ≥ 50 × 109/L at day 8, was observed in 65.6% of patients receiving avatrombopag versus 0% of patients receiving placebo (p < 0.0001). On the exploratory end point of durable response, attainment of durable response (i.e., platelet counts ≥ 50 × 109/L in ≥ 6 of the final 8 weeks of treatment during the core phase and no use of rescue therapy) was significantly more likely with avatrombopag versus placebo (34.4% vs 0.0%; p = 0.009) [6].

2.3.3 Eltrombopag

In a phase III study of eltrombopag, 197 patients were randomized in a 2:1 ratio to eltrombopag 50 mg daily or placebo. The mean number of weeks of platelet response reported in patients receiving eltrombopag was 14.3 weeks [7]. Although day 8 platelet response rates were not explicitly reported in the phase III trial of eltrombopag in ITP, median day 8 response rates with eltrombopag 50 mg and 75 mg were 44% and 62%, respectively, in an earlier phase II study of eltrombopag, all of which were higher than the < 10% rate of platelet response with placebo [38]. In a post-hoc analysis of available data, durable response (i.e., platelet counts ≥ 50 × 109/L in ≥ 6 of the final 8 weeks of treatment during the core phase regardless of prior rescue therapy use) occurred in 60% of patients (57/95) randomized to eltrombopag who completed the core study versus 10% (4/39) of patients randomized to placebo (p < 0.0001). Overall rates of attainment of durable platelet response were 51% (19/37) in splenectomized patients receiving eltrombopag and 66% (38/58) in patients receiving eltrombopag who had not undergone splenectomy [7].

2.3.4 Romiplostim

Romiplostim was evaluated in two parallel phase III trials of splenectomized (n = 63) and non-splenectomized ITP patients (n = 62). In each trial, patients were randomized to romiplostim or placebo in a 2:1 ratio. The 1-µg/kg starting dose of romiplostim was increased by 1–2 µg/kg every week or every 2 weeks depending on platelet counts, with a maximum allowed weekly dose of 15 µg/kg. The mean number of weeks of platelet response with romiplostim versus placebo was 13.8 versus 0.8 weeks (p < 0.0001) in the splenectomized group and 15.2 versus 0.8 weeks (p < 0.0001) in the non-splenectomized group. Early platelet response occurred in 25% of patients within 1 week of romiplostim initiation. The primary outcome measure of durable response (platelet counts ≥ 50 × 109/L in ≥6 of the final 8 weeks of treatment during the core phase and no use of rescue therapy during the prior 8 weeks) was achieved in 38% of splenectomized patients (16/42) receiving romiplostim versus 0% (0/21) of splenectomized patients receiving placebo (p = 0.0013). In non-splenectomized patients receiving romiplostim, the primary outcome of durable response was achieved in 61% of patients (25/41) versus 5% of patients (1/21) receiving placebo (p < 0.0001) [8].

2.4 Switching Between TPO-RAs

TPO-RAs differ in their molecular structure, pharmacokinetic profile, and the manner in which they stimulate the TPO receptor. Unlike romiplostim, avatrombopag and eltrombopag do not compete for the endogenous TPO receptor binding site [28, 39]. Binding to the TPO-receptor, c-Mpl, may induce greater Akt pathway activation compared with JAK/STAT, thereby inducing megakaryocytic proliferation to a greater degree than platelet production [39]. Although adherence to dietary restrictions while taking eltrombopag therapy may explain some differences in efficacy between TPO-RAs, subtle mechanistic differences appear to have clinically relevant effects, as patients who experience toxicities or lack of efficacy with one TPO-RA may benefit by switching to an alternate TPO-RA. In one study of 46 patients switching between TPO-RAs, the most common reasons for the switch included lack of efficacy of the first TPO-RA followed by platelet count fluctuation, side effects, and patient preference. Although suboptimal efficacy with an initial TPO-RA affects a minority of patients, pharmacists should note that patients who do not exhibit a response to one TPO-RA may exhibit a response to an alternative agent. For example, in romiplostim nonresponders, 46% of patients experienced a response in the platelet count after switching to eltrombopag. Similarly, 80% of eltrombopag nonresponders who switched to romiplostim responded to the new treatment [40].

A pooled analysis by González-Porras et al. of data from 18 publications identified lack of efficacy (defined by sustained platelet counts < 30,000/µL) as the reason for switching to another TPO-RA in 58% of patients. In patients who switched as a result of efficacy concerns, nearly two-thirds (65%) of patients who did not respond to a prior TPO-RA responded to an alternative TPO-RA. Among patients who switched for reasons unrelated to efficacy, 93% of patients maintained their response with the subsequent TPO-RA, with response defined as sustained platelet counts ≥ 30,000/µL. Overall, across trials, 78% of patients achieved or maintained a platelet response [39]. Outcomes were similar regardless of which TPO-RA was used initially, with a higher rate of response when switched due to adverse events or patient preference versus lack of efficacy [39, 41]. A 4-week trial at the maximum recommended dose is generally considered adequate before transitioning between TPO-RAs or adding on a second agent to augment efficacy (e.g., immunosuppressive therapy or corticosteroid) [39].

3 Treatment and Dispensing Considerations

3.1 ITP Treatment Considerations for Pharmacists

The updated American Society of Hematology (ASH) 2019 guidelines for ITP suggest second-line treatment with a TPO-RA, splenectomy, or rituximab in corticosteroid-dependent or corticosteroid-unresponsive adults with at least a 3-month history of ITP. Although the choice should be individualized, the guidelines generally recommend use of rituximab over splenectomy and treatment with a TPO-RA over rituximab [4]. Pharmacists should note that receipt of rituximab may diminish vaccine response for 6–12 months after receipt. It is therefore imperative that patients receive any vaccinations that are due prior to the administration of rituximab and consider the diminished response if patients will be travelling to an area where specific vaccines are recommended prior to travel [41, 42]. Thus, selection of treatments requires consideration of patient age, route of administration, diet, monitoring needs, and drug cost, and involves shared decision making by the patient, family, and physician. The pharmacist’s role in patient education and counseling is important for shared decision making while considering various aspects including the duration of ITP therapy, corresponding adverse effect profile, route of administration, frequency of bleeding episodes requiring hospitalization or rescue medication, patient age, comorbidities, medical and social support, patient preferences, cost, and availability [4, 5].

In this discussion, it is important to recognize that patients and physicians have different perceptions and priorities with regard to treatments in ITP. For example, although physicians perceive bleeding as the most important factor in quality of life, adult patients with ITP report that low energy levels and limitations on the ability to perform daily tasks were more salient. Nearly half (43%) of working patients reported that ITP affected their productivity at work, and approximately one-fifth (21%) reported considering termination of their employment due to ITP [43]. Effects on daily activities and quality of life are a generally under-reported concern. TPO-RAs have been shown to increase platelet counts and reduce bleeding. Interestingly, the most frequently cited reason for initiating a second-line therapy in children with ITP was quality of life (e.g., limitations in daily tasks and reduced energy levels) with 73% of clinicians reporting this as a reason to treat. Bleeding was less commonly cited as a factor in initiating therapy (15–18%) [44]. Pharmacists who understand how patient preferences affect the decision to initiate and continue specific types of therapy in ITP may wish to include these counseling points to inform patients of their options and help bridge this disconnect [43].

3.2 Administration Considerations

It is particularly important for pharmacists to understand and communicate restrictions on use of specific TPO-RAs with food. Administration of eltrombopag tablets with food versus in a fasted state reduced bioavailability as assessed by area under the concentration-time curve by 59% (geometric mean ratio [GMR]: 0.41; 90% confidence interval [CI] 0.36–0.46), and administration of eltrombopag tablets with aluminum hydroxide and magnesium carbonate antacid tablets reduced bioavailability by approximately 70% (GMR 0.30; 90% CI 0.24–0.36) [45]. When administered with a high-calcium meal or within 2 hours of an ordinary meal, the oral suspension formulation of eltrombopag showed reduced bioavailability versus the tablet formulation of eltrombopag. Specifically, least square mean GMR of area under the concentration-time curve was reduced by 75% (GMR 0.254; 90% CI 0.224–0.287) with the oral suspension versus the tablet formulation when administered with a high-calcium meal and by 20% (GMR 0.804; 90% CI 0.711–0.908) when administered 2 h before an ordinary meal [46]. The product labeling for eltrombopag recommends that doses of eltrombopag be taken either without a meal or with a meal containing 50 mg of calcium or less. Consistent with concerns over calcium–food interactions, eltrombopag should be taken either 2 hours before or 4 hours after any product containing polyvalent cations (e.g., calcium, iron, magnesium) [30]. Studies of avatrombopag bioavailability found no effect of food on the ultimate extent of avatrombopag absorption, and there are no restrictions on taking avatrombopag with polyvalent cations [28, 47]. The product labeling for avatrombopag recommends taking doses with food [28]. Although it is not currently indicated for use in ITP, lusutrombopag may be taken with or without food [29].

3.3 Dispensing Considerations

The only injectable TPO-RA, romiplostim, is supplied in vials containing 125 μg, 250 μg, or 500 μg. Although romiplostim is usually reconstituted and administered in a health care setting through subcutaneous injection [31], it is possible to procure this agent for home use. In this instance, extensive training must be performed to ensure that the adequate amount of diluent is used to dissolve the lyophilized powder and that the dosage is based on the final concentration as stated in the drug label, not based upon how much volume was used for reconstitution. Additionally, because volume sizes of one milliliter or less are common, great care must be taken to avoid pushing the plunger on the syringe so as to avoid loss of any product [31].

Oral TPO-RAs approved in ITP include eltrombopag and avatrombopag. Eltrombopag is available in tablets with four dosage strengths of 12.5 mg, 25 mg, 50 mg, and 75 mg, whereas the oral suspension is available in 12.5-mg and 25-mg packets to be reconstituted immediately prior to administration [30]. When dispensing eltrombopag, pharmacists should be aware that eltrombopag dosing may need to be reduced in patients who are of Asian ancestry (i.e., Chinese, Japanese, Taiwanese, or Korean) [30]. Unlike other TPO-RAs, avatrombopag is supplied only as 20-mg tablets for oral use. Because there is only one dosage strength for avatrombopag, downward titration of the dose may not immediately require a new prescription. Importantly, a new prescription is always required when titrating the dose to higher levels [28]. Note that tablets of avatrombopag should be stored in the original blister pack and eltrombopag should be dispensed in the original bottle. Current data are not robust enough to make a definitive recommendation on the use of TPO-RAs in pregnancy or lactation; benefits versus risks must be weighed for both mother and baby [48]. Dispensing information including associated toxicities is compiled in Table 2 [28,29,30,31, 37, 49,50,51,52,53,54,55,56,57].

Table 2 Indications and dispensing considerations of TPO-RAs [28,29,30,31, 37, 49,50,51,52,53,54,55,56,57]
Full size table

3.4 History of Risk Evaluation and Management System (REMS) Program

Risk Evaluation and Management System (REMS) exists to inform appropriate risk–benefit decisions to ensure safe use of specific medications. Although a REMS program existed for romiplostim and eltrombopag, this requirement was lifted for both products in December 2011. These risk management plans were in place to mitigate the risk of certain adverse events such as hepatotoxicity and bone marrow fibrosis, monitor the risk of hematologic malignancy progression in patients with myelodysplastic syndrome (MDS), and monitor for cases of hemorrhage resulting from worsening thrombocytopenia in patients with ITP [58]. Pharmacists should continue to counsel patients on the potential serious risks that may occur with TPO-RAs and refer patients to medication guides. A modified REMS program is still in effect that includes a communication plan to inform health care professionals about the potential for adverse events associated with treatment [59, 60].

3.5 Extemporaneous Compounding

Although avatrombopag and eltrombopag are administered orally, in some cases, compounding medications for oral administration in patients who cannot swallow an intact dosage form may be necessary. For these patients, eltrombopag is available as tablets and as an oral suspension. The eltrombopag oral suspension must be reconstituted with water prior to use (not hot water) and should be administered immediately and discarded if more than 30 minutes have elapsed since reconstitution [30]. It is currently not recommended to split, crush, or mix eltrombopag with food or liquids, as there are currently no published studies evaluating these types of manipulations [30, 45]. Stability data in food are available with avatrombopag. Finely crushed avatrombopag is stable for 30 minutes when mixed with whole milk yogurt or chocolate pudding. Recovery of avatrombopag in these foods was within 5% of the labeled amount, which is consistent with regulatory guidance on the administration of medications in a food vehicle [49].

3.6 Drug Interactions

TPO-RAs taken orally may be affected when co-administered with other drugs and require a dose adjustment. Patients must be advised to report use of any concomitant prescription or nonprescription medications or dietary supplements. Table 3 provides a list of the most common drug interactions observed [28,29,30,31].

Table 3 Drug interactions for TPO-RAs [28,29,30,31]
Full size table

3.7 Adverse Events with TPO-RAs in ITP

3.7.1 Avatrombopag—Common Adverse Events

TPO-RAs are generally tolerated well in patients with ITP. In a pooled analysis of clinical trials of avatrombopag in adult patients with chronic ITP, adverse events reported in ≥ 10% of patients included headache (31%), fatigue (28%), contusion (26%), epistaxis (19%), upper respiratory tract infection (15%), arthralgia (13%), gingival bleeding (13%), petechiae (11%), and nasopharyngitis (10%) [28].

3.7.2 Eltrombopag—Common Adverse Events

In adult patients with chronic ITP, the most common adverse reactions reported in at least 3% of patients were eltrombopag-included headache (10%), increased alanine aminotransferase (5%), increased aspartate aminotransferase (5%), cataract (5%), fatigue (5%), blood bilirubin increased (4%), nausea (4%), and diarrhea (3%) [30].

3.7.3 Romiplostim—Common Adverse Events

In two pooled trials of romiplostim in ITP, the most commonly reported side effects reported in ≥ 5% of patients were arthralgia (26%), myalgia (14%), pain in extremity (13%), shoulder pain (8%), dizziness (17%), paresthesia (6%), insomnia (16%), abdominal pain (11%), and dyspepsia (7%) [31].

The safety profiles of TPO-RAs are not superimposable. For example, adverse events such as cataracts and elevated transaminase levels are more commonly seen with eltrombopag. Certain potential adverse events, such as neutralizing antibody development, would only be possible with romiplostim [27].

3.7.4 Thromboembolic Events

Although ITP is associated with an elevated risk of thrombosis, thrombotic and thromboembolic complications are a possible adverse event with TPO-RAs, and patients treated with TPO-RAs should be monitored for thromboembolic events. Some long-term nonrandomized studies have suggested that patients with ITP have a modestly higher rate of thrombosis when treated with a TPO-RA than with an immunosuppressant [17]. The occurrence of thromboembolic events in long-term studies of avatrombopag, eltrombopag, and romiplostim in patients with ITP are comparable and have been estimated at approximately 6% to 7% across agents, with no clear association between platelet counts and risk of thrombosis [28, 30, 61,62,63]. According to international guidance, although TPO-RAs may be associated with arterial and venous thrombosis as a class warning, the rate of thromboembolic events was not elevated versus placebo in long-term (i.e., ≥ 6 month) clinical trials of avatrombopag or eltrombopag [5, 6, 62]. Of note, however, thromboembolic event risk may be elevated versus placebo in patients receiving romiplostim, particularly in patients 60 years or older [5, 63].

3.7.5 Reticulin Fibrosis

Data from a large number of patients treated with TPO-RAs demonstrate a low risk of generally reversible bone marrow reticulin fibrosis without an increased risk of collagen fibrosis [17]. The risk of progression to myeloid malignancies is also low in these patients. In most cases, the grade of fibrosis does not change with TPO-RA treatment, but a slight non-progressive reticulin fibrosis is observed in 10–50% of cases (i.e., grade 1 or Baumeister grade < 2). Severe grades of reticulin fibrosis are very rare across studies of TPO-RAs [27].

3.7.6 Transaminase Elevation

Transaminase elevation may occur in up to 10% of patients taking eltrombopag and is typically seen in the first year of treatment, with hyperbilirubinemia occurring in approximately 4% of patients with long-term use [27, 30]. Importantly, the US Food and Drug Administration (US FDA)–approved dose of eltrombopag for severe aplastic anemia, 150 mg daily for 6 months for patients 12 years and older, may cause skin and eye discoloration mimicking a jaundice-like state. Liver function tests along with iron indices should be examined to determine the cause. In a pooled analysis of four clinical studies of avatrombopag enrolling a total of 128 subjects, shifts in liver function tests to grade 2 or 3 abnormalities occurred in seven (5.5%) patients, with no cases of grade 4 abnormalities observed or any increases in bilirubin and no signal of avatrombopag-induced hepatotoxicity identified [64]. With the exception of cases of portal vein thrombosis in patients with chronic liver disease, romiplostim does not appear to be associated with liver function abnormalities [31]. Unlike eltrombopag, romiplostim and avatrombopag do not require monitoring of liver function tests during therapy [28,29,30,31].

3.7.7 Ocular Abnormalities

Cataracts have been reported with eltrombopag in patients with chronic ITP across three pooled trials but at an equal rate to that observed in the placebo group (7% vs 7%) [30]. In the case of romiplostim, an integrated safety analysis integrating data from 14 clinical studies identified 34 cases of cataracts (2.2 cases per 100 patient-years) in patients receiving romiplostim versus one case of cataracts reported in pooled control groups (0.9 cases per 100 patient-years) [27, 65]. Although cataracts have not been reported with avatrombopag, none of the TPO-RAs require regular monitoring for cataracts. Importantly, multiple confounding factors such as steroid use in patients with ITP preclude any firm conclusions about a potential relationship between use of any TPO-RA and cataracts in humans [27,28,29,30,31]. Adverse events of special interest with TPO-RAs across trials are summarized in Table 4 [27,28,29,30,31, 61, 64,65,66,67,68].

Table 4 Adverse events of TPO-RAs for ITP reported in clinical trials [27,28,29,30,31, 61, 64,65,66,67,68]
Full size table

4 Other Uses of TPO-RAs

4.1 TPO-RAs in Chronic Liver Disease and Perioperative Use

Avatrombopag and lusutrombopag are approved for the treatment of thrombocytopenia in adults with chronic liver disease who are scheduled to undergo a procedure [28, 29]. Avatrombopag was evaluated in the ADAPT-1 and ADAPT 2 studies. In both studies, patients were assessed for eligibility during the 2 weeks prior to study randomization and were assigned to cohorts of high or low baseline platelet counts. Patients were randomized in a 2:1 ratio to receive avatrombopag 40 mg daily for 5 days in patients with high baseline platelet counts (i.e., 40 × 109/L to <50 × 109/L) and 60 mg daily for 5 days in patients with low baseline platelet counts (i.e., < 40 × 109/L) or a matching placebo. The procedure was scheduled to take place 10–13 days after the initiation of avatrombopag (5–8 days after the last dose) [9, 28]. The primary end point in both studies was the proportion of patients who did not require a platelet transfusion or a rescue procedure to manage bleeding after randomization and for up to 7 days following an elective procedure [9].

In patients with low baseline platelet counts, a significantly higher rate of attainment of the primary end point occurred in the avatrombopag patients versus placebo in ADAPT-1 (66% vs 23%; p < 0.0001) and ADAPT-2 (69% vs 35%; p < 0.0006). Of note, according to these data, roughly one out of three patients may not meet the platelet goal, which may prevent them from moving forward with the planned procedure or may require platelet transfusion to proceed. In patients with high baseline platelet counts, a significantly higher rate of attainment of the primary end point was observed in the avatrombopag patients versus placebo ADAPT-1 (88% vs 38%; p < 0.0001) and ADAPT-2 (88% vs 33%; p < 0.0001) [9].

According to study authors, treatment-emergent adverse events in both ADAPT-1 and ADAPT-2 occurred at a rate similar to that observed with placebo [9]. Overall rates of treatment-emergent adverse events in ADAPT-1 were lower with avatrombopag versus placebo in patients with low baseline platelet counts (59.6% vs 64.6%) and high baseline platelet counts (53.4% vs 56.3%). Likewise, in ADAPT-2, rates of treatment-emergent adverse events were similar with avatrombopag and placebo in patients with low baseline platelet counts (51.4% vs 51.2%) and high baseline platelet counts (49.1% vs 45.5%) [9]. In a pooled analysis of ADAPT-1 and ADAPT-2, the most commonly reported treatment-emergent adverse events were pyrexia, abdominal pain, nausea, headache, upper abdominal pain, and procedural pain [69].

Lusutrombopag was evaluated in two placebo-controlled trials, L-PLUS-1 and L-PLUS-2, in which patients were treated with 3 mg of lusutrombopag for up to 7 days prior to an invasive procedure that occurred 2–7 days after completion of the course of lusutrombopag [10, 11, 29]. At baseline, per inclusion criteria, patients had a baseline platelet count < 50 × 109/L [10, 11]. In the L-PLUS-1 trial, patients with chronic liver disease with a planned invasive procedure were randomized in a 1:1 ratio to receive lusutrombopag 3 mg or placebo. The primary end point of the proportion of patients who did not require a platelet transfusion prior to their procedure was significantly higher with lusutrombopag versus placebo (79.2% vs 12.5%) [10]. In the L-PLUS-2 trial, which had a similar design to L-PLUS-1, the percentage of patients who did not require platelet transfusion prior to the procedure was higher in patients receiving lusutrombopag versus placebo (64.8% vs 29.0%; p < 0.0001) [11].

In the L-PLUS-1 trial, adverse drug reactions that may have been related to study drug occurred in a slightly higher percentage of patients receiving lusutrombopag versus placebo (8.3% vs 2.1%) [10]. However, rates of treatment-emergent adverse events in the L-PLUS-2 trial were similar in the lusutrombopag and placebo arms (47.7% vs 48.6%) [11]. Common adverse events occurring in > 3% of patients in L-PLUS-2 included headache, abdominal pain, fatigue, peripheral edema, and nausea [11], whereas the most common adverse events observed with lusutrombopag and placebo in L-PLUS-1 were postoperative fever, procedural pain, procedural hypertension, and increased aspartate aminotransferase [10]. Importantly, as with all TPO-RAs, it is important for patients to be monitored for post-operative thromboembolic complications [28, 29].

Some evidence supports use of TPO-RAs prior to other medical procedures when platelet transfusions are in limited supply and provided there is adequate lead time. In a recent case report by Lim and Gilreath published in 2020, periprocedural use of avatrombopag eliminated the need for pre-operative platelet transfusion in a 63-year-old man with chronic liver disease who was undergoing spinal decompression surgery. Because certain neurosurgical procedures require that patients meet a minimum platelet threshold (e.g., 100 × 109/L), TPO-RAs have the potential to reduce the need for blood products. The patient initiated avatrombopag 20 mg daily starting 4 weeks before surgery. After 2 weeks, when the platelet count target was not achieved, the dose was increased to 40 mg three times weekly and 20 mg daily administered on the remaining 4 days of each week. Avatrombopag raised platelet levels from 63 × 109/L to 100 × 109/L by the time of surgical intervention [70].

4.2 Severe Aplastic Anemia

Eltrombopag has been approved in combination with standard immunosuppressive therapy for the first-line treatment of adult and pediatric patients aged 2 years and older with severe aplastic anemia (SAA) [30]. Eltrombopag at a dose of 150 mg daily combined with immunosuppressive therapy, consisting of a 4-day course of equine antithymocyte globulin and cyclosporine (based on ideal body weight), was evaluated in a phase I/II investigator-initiated trial evaluating three cohorts of patients with SAA with differing times to initiation of combined immunosuppressive therapy following onset of SAA (i.e., day 14 to month 6 in cohort 1 [n = 30], day 14 to month 3 in cohort 2 [n = 31], and day 1 to month 6 in cohort 3 [n = 31]). Patients were evaluated for hematologic response, including complete response (i.e., absolute neutrophil count ≥1000 per microliter, a hemoglobin level ≥10 g/dL, and a platelet count ≥ 100 × 109/L) and partial response (i.e., no longer meeting SAA diagnostic criteria but without meeting all three complete response criteria). Complete response by month 6 occurred in 58% of patients in cohort 3, 26% of patients in cohort 2, and 33% of patients in cohort 1. Notably, the overall response rate (complete and partial response) at month 6 in cohort 3 was 80%, which was significantly higher than the 66% overall response rate observed in a historical cohort of 102 patients with SAA who received antithymocyte globulin and cyclosporine alone [71]. Although eltrombopag is approved for use in combination with standard immunosuppressive therapy in adults and children aged 2 years and older with severe aplastic anemia, a recent subgroup analysis of 40 pediatric patients receiving eltrombopag plus immunosuppressive therapy versus a historical cohort of 87 children receiving immunosuppressive therapy alone did not identify a significant improvement in overall response rates (70% with eltrombopag plus immunosuppressive therapy versus 72% in the historical cohort of children receiving immunosuppressive therapy alone; p = 0.78) [30, 72].

The most common adverse events occurring in ≥ 20% of patients treated with eltrombopag in an open-label clinical trial of adults with refractory SAA were nausea, fatigue, cough, diarrhea, and headache [30]. When combined with antithymocyte globulin and cyclosporine, severe adverse events attributed to eltrombopag in 92 patients with SAA included liver test abnormalities (18%), abdominal pain (2%), maculopapular rash (2%), and pruritus (1%) [71]. Importantly, patients taking ≥ 150 mg/day should be cautioned that skin and eye hyperpigmentation may occur, owing to either iron-chelation properties of eltrombopag leading to a transient, drug-dependent, hemochromatosis-like effect, or due to the drug distribution and discoloring effect of the molecule itself in these various tissues [73].

4.3 Myelodysplastic Syndrome

Studies into the efficacy of TPO-RAs in MDS failed to show a survival benefit and indicated some potential for harm as a result of a potential increase in progression to leukemia [27]. The Study of Eltrombopag in Myelodysplastic Syndromes Receiving Azacitidine (SUPPORT) phase III trial did not identify a progression-free survival or overall survival benefit with eltrombopag plus azacitidine versus placebo plus azacitidine. Although the risk of death or progression to acute myeloid leukemia was not elevated with combination therapy, the risk of serious adverse events and adverse events leading to discontinuation were higher with combination therapy. It is important to alert prescribers of this combination to the potential for increased risk of adverse events without added clinical benefit [74].

4.4 Post-Hematopoietic Stem-Cell Transplant

Persistent isolated thrombocytopenia often occurs after allogeneic hematopoietic stem-cell transplantation (allogeneic SCT) and may increase the risk of adverse transplant outcomes. Secondary failure of platelet recovery (SFPR) may be defined as platelet counts < 20 × 109/L for 7 days consecutively or a need for transfusion in patients with a platelet count ≥ 50 × 109/L who did not require transfusion in the 7 days following SCT. Although there is no well-established standard of care in SFPR management, patients may receive supportive care. Platelet transfusion may be used in some cases; however, adverse events may occur, such as fluid overload, infusion-related reactions, infection, or refractoriness to subsequent platelet transfusions. TPO-RAs have been used as off-label treatment for these conditions with variable results [75]. Table 5 provides efficacy and safety data from a compilation of various clinical trials and meta-analyses using these treatments [75,76,77,78,79].

Table 5 Compilation of post-HSCT clinical trials and meta-analyses [75,76,77,78,79]
Full size table

4.5 Chemotherapy-Induced Thrombocytopenia

Chemotherapy-induced thrombocytopenia (CIT) is a serious complication of chemotherapy in cancer patients resulting in low platelet counts that may necessitate the reduction of chemotherapy doses, cycle delays, or a change in chemotherapy regimen, all of which may compromise long-term outcomes. In December 2019, the US FDA granted avatrombopag Orphan Drug Designation for the treatment of patients with CIT [80, 81].

In a phase III trial of avatrombopag to study its efficacy in CIT (NCT03471078), avatrombopag increased platelet counts as anticipated and as demonstrated in other populations of patients experiencing thrombocytopenia; however, the placebo group performed unexpectedly well with strong platelet count rebound prior to chemotherapeutic dose and few patients requiring platelet transfusions or chemotherapy dose modifications. Although initial results of the trial were negative on the primary composite outcome (i.e., the proportion of subjects who did not require platelet transfusion, a ≥ 15% reduction in the dose of chemotherapy, or delay of chemotherapy administration by ≥ 4 days), of note, the study excluded patients with grade 2 or higher chemotherapy-induced thrombocytopenia (other than the current chemotherapy regimen) within 6 months of screening and patients who had received more than two previous lines of chemotherapy. Moreover, the study evaluated patients with isolated, nadir thrombocytopenia and did not separately evaluate patients with more severe or persistent CIT, such as CIT that continued despite adequate time for recovery of white blood cell counts [80, 82].

Current treatment options for CIT are limited [83]. However, some data support the efficacy of TPO-RAs in CIT. A retrospective study published by Al-Samkari et al. in 2020 evaluated use of romiplostim for CIT across four centers. In this study, patients were evaluated on a primary outcome of attainment of platelet counts ≥ 75 × 109/L, with an increase of ≥ 30 × 109/L from baseline levels. A total of 173 patients with cancer were included in the study, including 153 patients with solid tumors. Over a median of four cycles (range: 1–36 cycles) of chemotherapy administered concurrently with romiplostim, response consistent with the primary efficacy outcome occurred in 71% of patients. Furthermore, 79% of patients did not require chemotherapy dose reductions or treatment delays, and 89% of patients did not require platelet transfusion [84].

Further studies of TPO-RAs in CIT are ongoing, including a study of avatrombopag in patients with any of several malignant tumor types (NCT04609891) [85], a study of avatrombopag in patients with thrombocytopenia after a hematopoietic stem-cell transplant (NCT04312789) [86], a study of romiplostim for CIT in pediatric patients with solid tumors undergoing myelosuppressive chemotherapy (NCT04671901) [87], and a study of romiplostim in patients with lymphoma receiving standard chemotherapy for lymphoma (NCT04673266) [88]. Romiplostim for CIT management is also being studied in non–small-cell lung cancer and gynecologic malignancies (NCT03937154) [89] as well as in gastrointestinal malignancies (NCT03362177) [90]. The optimal population for potential use of TPO-RAs in CIT requires further study.

5 Conclusions

In evaluating the characteristics of TPO-RAs, pharmacists should take into account not only efficacy and safety data but also patient preferences for route of administration, the need for stable platelet count control, and whether dietary restrictions pose an issue. Although strong data support the use of TPO-RAs in ITP, chronic liver disease, and a variety of other disease states, taking into account the pharmacokinetic, pharmacodynamic, and mechanistic characteristics of these agents is important. Analyzing the onset of action based on available data may also be relevant in therapeutic decisions. In addition, the cost of therapy may be taken into account across the useful dosage range of available agents. Pharmacists should counsel patients on the risk of thromboembolic complications and the relative risk of adverse events of special interest with TPO-RAs such as liver enzyme elevations. Through an in-depth knowledge of the TPO-RA class, pharmacists can better serve patients and other health care professionals, aiding in appropriate decisions across health systems and for individual patients.