Eltrombopag (Promacta) is the first orally absorbed, small-molecule, thrombopoietin receptor (TPO-R) agonist, approved (on November 20, 2008) by the US Food and Drug Administration (FDA) for the treatment of chronic immune thrombocytopenia (ITP) in patients who have relapsed following treatment with corticosteroids, immunoglobulins, and/or splenectomy. It joins romiplostim (NPlate) as a second thrombopoietic agent licensed by the FDA for chronic ITP. The Oncology Drug Advisory Committee (ODAC) meeting, where the data commented on by Dmytrijuk was presented, was held on May 31, 2008. At the time of this writing, that was 17 months ago, and the clinical development of eltrombopag has advanced. This commentary will provide an update since that time.
FDA approval was based on two 6-week, double-blind, randomized, placebo-controlled clinical trials and the initial results of a single-arm, open-label extension study. The two 6-week studies showed that eltrombopag treatment resulted in dose-dependent platelet count increases in most patients within 1 to 2 weeks regardless of splenectomy status, concomitant (stable) ITP medication (eg, prednisone(Drug information on prednisone)), or baseline platelet counts.[1,2] Treated patients were also more likely to experience platelet counts > 200,000/µL; if this occurred they were withdrawn from further treatment to avoid risks of thrombocytosis.[1,2] Dmytrijuk and coauthors also comment that the reduction in bleeding in these first two trials (using the World Health Organization scale) was only in the mildest category of bleeding (bruises and petechaie, grade 1) and therefore not clinically important.
Subsequently a worldwide, 197-patient, randomized, 6-month placebo-controlled trial was performed that was similar to the pivotal trials of romiplostim[4,5] in a number of ways (but not in others). This trial had similar entry criteria (eg, platelet count less than 30,000/µL, ITP of at least 6 months duration, and use of at least one previous treatment) and also had a 2:1 active drug–to-placebo ratio. The investigators reported a highly significant odds ratio (> 8) favoring patients on eltrombopag for a response to a count > 50,000/µL at any point during the trial. In this larger, longer study there was not only a significant reduction in overall bleeding (grade 1–4) but also a significant reduction when only grade 2–4 bleeding was considered. This reduction in bleeding depends on patients having a platelet response to the agent. Even when the three studies are considered together, the data on hemostatic risk (eg, tolerance of invasive procedures) are available for too few patients to comment upon, but there is no indication that eltrombopag-induced platelets are dysfunctional.
As of the FDA review, 109 patients had received eltrombopag in the Eltrombopag Extended Dosing Study (EXTEND). Currently, 299 patients have been treated in the EXTEND study, with 48 receiving eltrombopag for more than 18 months. More than 85% of patients have achieved platelet counts ≥ 50,000/µL and maintained median platelet counts ≥ 50,000/µL for most of the study period. This consistent response is similar to findings with romiplostim.
The safety data showed a risk for hepatotoxicity (see Table 1). Further analysis of safety data across the clinical program showed that the frequency of liver toxicities was similar in patients treated with eltrombopag or placebo except for alanine transferase (ALT) elevations. Approximately 5% of eltrombopag-treated patients experienced ALT ≥ 3 × the upper limit of normal, compared with 2% of placebo-treated patients. Approximately 2% of patients withdrew from the study due to elevations in ALT or bilirubin. Liver toxicities were typically mild, reversible, and unaccompanied by clinical symptoms; some patients had more marked elevations and discontinued therapy. As stated by Dmytrijuk, it remains important to monitor liver function tests every 4 weeks or so, even on long-term eltrombopag.
The potential for serious hemorrhage and worsened thrombocytopenia following discontinuation of eltrombopag were also identified as potential safety issues, with certain patients experiencing platelet counts < 10,000/µL. The true incidence of this is very difficult to determine precisely because the starting platelet count is very low in most patients; therefore, the incidence remains unknown. In a recent analysis of safety data from the ITP program, 8% of patients treated with eltrombopag (n = 20) and 8% of patients treated with placebo (n = 10) experienced decreases in platelet counts to < 10,000/µL and more than 10,000/µL below baseline platelet counts in the 4 weeks following discontinuation of study medication. None of the 10 placebo-treated patients and 3 of the 20 eltrombopag-treated patients had bleeding events associated with these decreases in platelet counts. This remains a concern for all thrombopoietic agents and requires monitoring especially in the second week following discontinuation. Thus far, induction of malignancy was only seen with romiplostim in patients with myelodysplastic syndromes (MDS). However, no published data have described the use of eltrombopag in patients with MDS or in chemotherapy-induced thrombocytopenia.
Another important area that is very slowly coming into focus with this class of agents is the risk of thromboembolic events (TEE). All of the double-blind, placebo-controlled studies with both of these agents have failed to demonstrate a statistically significantly increased risk of TEE. Nevertheless, data collection across all ITP studies suggested that there may be an increased risk of TEE in patients on eltrombopag. TEE appears to be a class effect, due in large part to the substantial increases in platelet count seen in treated patients and the prothrombotic tendency of ITP in a fraction of patients. It could be speculated that TEE is in fact not more frequent because of the presumed antiapoptotic effects of these agents on platelets and the potential role of apoptosis in platelet activation and thrombosis.
Finally, thus far increased reticulin deposition in the marrow has only been reported with romiplostim, although preliminary data demonstrating this effect with eltrombopag were presented at the ODAC meeting. The incidence and clinical significance of reticulin deposition—and whether any treated patients will progress to more significant fibrosis—are currently unknown. A condition of licensure for both agents is to perform studies in this area.
Although Dmytrijuk and coauthors do not comment on this, there is little pediatric data for these drugs at this time. One 22-patient study with romiplostim will be presented at the 2009 annual meeting of the American Society of Hematology, and a larger pediatric study with eltrombopag is just beginning. Additional evaluation of safety and efficacy is underway for both agents. The use of a Risk Evaluation and Mitigation Strategy (REMS), commented upon by Dmytrijuk and colleagues, has the major advantage of allowing this novel and highly effective class of agents to be available to patients while simultaneously monitoring substantially larger numbers of patients to better determine adverse event rates.