Clinical outcomes in eight patients with immune thrombocytopenia each treated with the three approved thrombopoietin receptor agonists

Immune thrombocytopenia (ITP) is an acquired autoimmune disorder characterized by isolated thrombocytopenia <100 × 109/L due to accelerated platelet destruction and/or impaired platelet production.1 Currently in the United States, there are three FDA-approved thrombopoietin receptor agonists (TPO-RA) for chronic ITP: eltrombopag, romiplostim and avatrombopag. All three have high response rates of 40%–90% depending upon the response definition and patient population.2 Previous studies demonstrated that sequential use of romiplostim or eltrombopag in either order resulted in a high response rate for the second agent. This is particularly true if there was a response to the first TPO-RA, for example, if the response was lost or if the patient was switched for a non-efficacy issue like toxicity. Lower response rates were observed in patients who did not respond to the first TPO-RA.3-5 Data comparing use of all three TPO-RAs during a patient's treatment course are limited. Between July 2019 (following FDA approval of avatrombopag) and September 2020, there were eight patients seen at New York Presbyterian Hospital – Weill Cornell (NYP-WC) with chronic ITP who had been treated with all three TPO-RAs. De-identified clinical data were collected from electronic medical records including demographics, duration of ITP, ITP treatment, response to and tolerance to TPO-RAs. The study was approved by the Institutional Review Board of NYP-WC with waiver of informed consent. Response was defined as platelet count ≥30 × 109/L and doubling of the baseline count with absence of bleeding without concomitant administration of a first line therapy. Statistical analysis was limited to descriptive measures. The eight patients (five females, three males), median age 64 (range 33–89) all had primary ITP with use of median four (range 2–13) ITP medications prior to starting a TPO-RA (Table 1). All eight had difficult to treat ITP with episodes of severe thrombocytopenia despite chronic treatment, periodically requiring rescue with intravenous immune globulin (IVIG) and/or corticosteroids. Three patients started their first TPO-RA at another institution (1, 2, 6). Initial TPO-RA choice was influenced by availability, insurance, patient preference, and financial limitations. Four patients received eltrombopag followed by romiplostim while four started with romiplostim followed by eltrombopag. All eight received avatrombopag last due to its being the most-recently licensed of the three agents. No patient went directly from one TPO-RA to another without intervening treatment and all received other treatments in conjunction with or in between TPO-RAs. Seven of eight patients responded to initial exposure to their first TPO-RA. Among the four patients who first received eltrombopag, three responded with median 12 months on treatment. The shortest was 1 week (allergic reaction) and the longest 8 years. One patient (7) was successfully tapered off eltrombopag until ITP relapse one year later (Table 1). Patient number eight did not respond as she received IVIg every 1–2 weeks during her initial 7 months on eltrombopag. However, this patient chose to re-initiate eltrombopag at the same dose (75 mg) after several weeks off and did not require IVIg for almost 3 months. The four patients who first received romiplostim all responded with a median 18 months on treatment (range 4–96 months). Two (1, 6) however, stopped treatment after 4 months due to unstable platelet responses and need for rescue therapy (Table 1). Six of eight patients (75%) responded to their second TPO-RA. For the four who initially received eltrombopag, all responded to romiplostim with median time on treatment 36.5 months (range 21–48 months). Reasons prompting the switch from eltrombopag to romiplostim included: no response/lost response (2), allergic reaction (1), and insurance/financial issues (1) (Table 1). Two of the four patients who received romiplostim first responded to eltrombopag with median time on treatment 6.5 months. Notably, one non-responder (patient 1) received only 1.5 weeks of eltrombopag without an increase to maximum dose; details are unclear as this occurred at an outside institution. Prior to availability of avatrombopag, six patients returned to a TPO-RA they had previously received. Five required higher doses on re-exposure, most without robust response (patients 1, 3, 5, 6, 7; Table 1). Four of eight patients underwent bone marrow biopsy prior to starting avatrombopag; None revealed evidence of an underlying hematologic malignancy and two had stable bone marrow fibrosis (MF-1). Six of eight (75%) patients responded to avatrombopag. Reasons prompting the switch to avatrombopag from the previous TPO-RA were patient preference (n = 4), lost response (n = 3), no response (n = 1). As of this report, five remain on avatrombopag with responses ranging from 8 to 16 months duration (Table 1). One patient (8) responded for 4 weeks. Two patients (2, 5) started prednisone, initially at a dose of 20–60 mg, concomitantly with the preceding TPO-RA a few weeks prior to starting avatrombopag; both required resumption of steroids for platelets <30 × 109/L after 8 and 9 months on avatrombopag, respectively (Table 1). The other three patients (7, 4, 6) remain on single agent avatrombopag. The two patients who did not respond to avatrombopag, also did not respond to eltrombopag. TPO-RAs represent an important treatment modality for patients with ITP. The high response rates and prolonged ability to safely use these agents has resulted in improved health and quality of life for many patients. While all three agents confer high response rates, no direct clinical comparisons exist. There are differences between the three agents in mechanism of action. Romiplostim binds to the TPO receptor at the endogenous TPO binding site, whereas eltrombopag and avatrombopag bind to a transmembrane portion of the receptor. The latter approach could yield improved responses through the additive effects of native TPO in combination with the TPO-RA.6 Furthermore, although eltrombopag and avatrombopag both bind to the transmembrane part of the TPO receptor, eltrombopag's effect is dependent upon its ability to chelate intracellular iron whereas avatrombopag is not a chelator.7 This difference explains their different dietary requirements – eltrombopag should be taken on an empty stomach while it is suggested that avatrombopag be taken with food. One study of endogenous TPO levels compared eltrombopag and romiplostim and suggested that romiplostim was effective in ITP patients at higher endogenous TPO levels than eltrombopag suggesting that romiplostim was a stronger agent.8 The TPO levels were not measured in this study. All eight of our difficult to treat ITP patients responded to at least one TPO-RA. Response rates of 75% were observed to a second, and subsequently third, TPO-RA. In this small sample, patients who did not initially respond to eltrombopag tended to lack response to avatrombopag. Among responders to each TPO-RA, there were patients who switched to a different TPO-RA due to lack of efficacy of a preceding TPO-RA as well as issues not related to efficacy (preference for an oral agent, intolerance, or insurance/financial). Re-exposure to a previously-used TPO-RA resulted in requirement of higher doses in five of six patients and also less robust response in three patients. Reasons for this lesser response are unclear but could represent evolution of clonal T cells.9 As an example, one patient on romiplostim had steadily tapered his dose down from 15 to 7 μg/kg/week. After trying eltrombopag without success, he no longer responded to romiplostim on re-initiation. Notwithstanding heterogeneity for interval of time between TPO-RAs, dose of TPO-RAs, and use concomitant ITP treatment in between and in conjunction with TPO-RAs, we believe that the described reflects" real-world" experience with ITP. There are several limitations in addition to small sample size. We did not assess TPO levels or TCR clonality. Not all patients received the maximum dose or adequate time on all TPO-RAs. Despite these limitations, these data support considering these three TPO-RAs as separate treatment options given good responses to sequential treatment. An exception might be that a lack of response to eltrombopag may predict a similar lack of response to avatrombopag. This would be consistent with both agents binding in the same transmembrane region of the TPO-R, distinct from that of endogenous TPO or romiplostim. Although combinations of TPO-RAs were not used in the current study, there are now preliminary results of a randomized controlled trial suggesting safety and efficacy of this approach.10 In conclusion, we would recommend, particularly for patients who have highly refractory ITP, to be cautious about switching from one TPO-RA to another unless due to toxicity or loss of response. Not only might the patient not respond to the other TPO-RA but there might be a diminished response on re-trial of the original agent. Similarly, initial lack of or loss of response to one TPO-RA should not preclude trying another but it appears better to switch between romiplostim and either eltrombopag or avatrombopag than to switch between eltrombopag and avatrombopag. These recommendations are preliminary and further investigation is encouraged. Eun-Ju Lee has served on the advisory board for Principia. Madhav Seshadri has no disclosures. James B. Bussel has served on advisory boards and/or consulted for Amgen, Novartis, Dova, Rigel, UCB, Argenx, Momenta, Regeneron, RallyBio, and CSL-Behring. Eun-Ju Lee, Madhav Seshadri and James B. Bussel contributed to the data acquisition and interpretation of data. Eun-Ju Lee and James B. Bussel wrote the manuscript. All authors provided input on the manuscript and approved the final version for submission. The data that support the findings of this study are available from the corresponding author upon reasonable request.