Efficacy and Long-Term Outcomes of Eltrombopag Treatment Within 6 Months of Diagnosis in Patients With Steroid Unresponsive or Dependent Immune Thrombocytopenia

Abstract

Immune thrombocytopenia (ITP) is an autoimmune disorder associated with the production of autoantibodies directed against platelet glycoprotein complexes, resulting in both accelerated destruction of platelets and a relative reduction in platelet production. In adults, the majority of the acute cases respond to initial steroid therapy, but 50% relapse by 6 months and an additional 25% relapse beyond 1 year [1]. The non-responders and early relapsers can be difficult to manage and require additional therapies. These include intravenous immunoglobulin (IVIg), splenectomy, and rituximab, all of which can be problematic with variable and non-sustained responses, and specific thrombotic and infective risks with splenectomy [2].

Thrombopoietin receptor agonists (TPO-RA) exert their effects in ITP by increasing platelet production. Clinical trials have shown that TPO-RA are efficacious and well tolerated in chronic ITP patients, with some patients achieving a lasting response after cessation [3]. At the time, the study described in this manuscript was designed (2014), little was known about the short- and long-term response to TPO-RA in patients with newly diagnosed ITP. Hence, this study was initiated to evaluate the early- and long-term efficacy and safety of eltrombopag in adult non-splenectomized patients with steroid-unresponsive or dependent ITP within 6 months of diagnosis.

Patients at six Australian centers were enrolled into this multicenter, single arm, prospective open-label study between February 2014 and February 2016. Patients were 18 years or older and with ITP diagnosed within the prior 6 months which was steroid-unresponsive or -dependent and, if used, without a sustained response to IVIg. Full inclusion and exclusion criteria can be found in the Supporting Information. The study was conducted following approval from the ethics committees of participating institutions and patients provided written informed consent. The study was registered as a clinical trial [ACTRN12613000721707].

Patients with platelets < 10 × 10⁹/L commenced 75 mg eltrombopag daily; those with a platelet count ≥ 10 × 10⁹/L commenced 50 mg daily. The dose was progressively increased by 25 mg increment every 2 weeks to maximum of 150 mg daily (100 mg daily if East Asian) if the platelet count remained ≤ 30 × 10⁹/L or there was clinically significant bleeding. The dose in responding patients was maintained until the week 12 response assessment, unless the platelet count was ≥ 200 × 10⁹/L. Subsequently, the dose was adjusted to the minimum required to maintain a platelet count > 30 × 10⁹/L but ideally > 50 × 10⁹/L, preferably in the absence of concomitant steroids. After 6 months and beyond, patients who achieved at least a stable partial response (PR; platelets > 50 × 10⁹/L) had eltrombopag tapered by 25 mg each 3 weeks, with resumption of the previous therapeutic dose if the platelet count fell below 30 × 10⁹/L.

Response was assessed in all patients weekly until week 12, then at further time points to week 52 or 130 (see Study schema in Supporting Information). The primary endpoint was the overall response rate (ORR) at week 12: the composite of patients achieving a complete response (CR; platelet > 100 × 10⁹/L), PR (platelets 50–100 × 10⁹/L) or minor response (MR; platelet 30–49 × 10⁹/L with ≥ 50% reduction in the dose intensity of concomitant ITP therapy compared with screening). The protocol specified if a PR or greater was reached at week 12 then the platelet count should be confirmed at week 14–16. However, after noting eight patients did not have a confirmatory platelet count the analysis was modified to allow inclusion of these patients in the response outcome assessment; this is subsequently referred to as “amended analysis.” Secondary endpoints included CR, PR and MR rates at weeks 4, 8, and 26, therapeutic response at week 26 and 130, and safety profile. Although not specified in the original study design, long-term follow up of participants beyond week 130 was conducted to assess long-term responses and ITP management.

Thirty-nine patients were enrolled. The median age was 52 years (range 18–82), 18 (46%) were female and the median ITP duration at enrolment was 2.2 months (range 0.3–6). Prior therapies apart from prednisolone included IVIg in 27 (69%) and azathioprine in 10 (26%) patients.

The median platelet count at baseline was 21 × 10⁹/L (range 2–94) and at week 12 was 117 × 10⁹/L (range 20–498). Figure 1 outlines responses at each time point. The ORR per protocol and per amended analysis at week 12 was 64% (90% CI 52–77; CR 41%, PR 15%, MR 8%) and 85% (90% CI 75–94) (CR 59%, PR 18%, and MR 8%), respectively. Platelet counts at week 12 in those who achieved an MR or greater were 127 × 10⁹/L (range 58–498) and 142 × 10⁹/L (range 42–377). The median eltrombopag dose at weeks 12 and 27 was 50 mg, and at weeks 52, 104, and 130 it was 25 mg. Of the 29 patients who completed the study at week 130, 12 (31% of the total study population) had ceased eltrombopag and remained in CR.

The median prednisolone dose at baseline was 25 mg (range, 5–100) and at week 12 was 2.25 mg (range, 0–15). Of the 23 patients who achieved a CR at week 12, 17 had ceased steroids and six remained on steroids at a median dose of 5 mg (range 5–15). Six required rescue therapies with steroids (6) and IVIg (2) for an ITP relapse throughout the study. Four of these participants subsequently came off study; the other two completed study and remained on eltrombopag with ongoing CR.

With long-term follow up (median of 83.5 months from enrolment), 16 patients (41%) had stopped all ITP therapies and maintained a MR or better. Twelve (31%) remained on eltrombopag alone (≤ 50 mg daily) with ongoing PR or better. Two (5%) remained on high dose eltrombopag (150 mg daily) with additional therapies (prednisolone, hydroxychloroquine) to maintain response. The remaining nine (23%) patients stopped eltrombopag, at time points indicated in Figure 1 and required alternative therapies.

A symptomatic adverse event occurred in 77% of participants. The most common were headache (28%), nausea (21%), musculoskeletal pain (15%), rash (15%), and diarrhea (15%). These were mild (grade 1–2) except for one grade 3 headache. Two patients had grade 2 transaminitis, who were on doses of 150 and 125 mg, respectively, which necessitated eltrombopag dose reduction. Serious AEs (SAEs) occurred in 9 patients (23%) of which three were considered treatment-related: one grade 3 transaminitis (while on a dose of 125 mg) which resolved with eltrombopag cessation and two venous thromboembolism (VTE)—deep venous thrombosis in a 76 year-old male, at a platelet count of 97 × 10⁹/L receiving 125 mg eltrombopag, which was subsequently ceased, and pulmonary embolism in a 71 year-old female with concurrent ANCA-positive vasculitis, with a platelet count of 230 × 10⁹/L receiving 50 mg eltrombopag which was continued. Bleeding occurred in 14 (36%) patients, most (9; 64%) were grade 1. Grade 2 bleeding occurred in four patients: menorrhagia; gastrointestinal bleed requiring endoscopy in a patient on aspirin with a platelet count > 100 × 10⁹/L; haematuria; and epistaxis. There was one grade 3 gastrointestinal bleeding in a non-responding patient early in the study. Four patients had thrombocytosis (platelet count ≥ 450 × 10⁹/L), which resolved with eltrombopag dose reductions without clinical sequelae.

This study is unique in describing both the efficacy and tolerability of eltrombopag in non-splenectomized adults with steroid unresponsive or dependent ITP within 6 months of diagnosis, as well as the natural history with long-term follow up. We observed a high initial response rate with 85% of participants achieving a response at week 12, and most able to wean prednisolone completely or to small doses. Most responses were durable with over 70% maintaining response (40% off all therapy and 30% on eltrombopag) at both week 130 and with follow up at 7 years, a sustained response rate similar to that seen with eltrombopag in chronic ITP [4].

The few studies evaluating eltrombopag in newly-diagnosed ITP have reported similar outcomes but with short follow-up. A single-center study of 25 patients with steroid-nonresponsive newly-diagnosed ITP, found 76% had a sustained response to eltrombopag at 3 months [5]. Retrospective and real-world studies confirm that response to eltrombopag is similar between newly diagnosed, persistent, and chronic ITP patients [6]. It is likely that the ability to maintain sustained responses represents the natural history of ITP rather than a unique effect of TPO-RAs as similar trajectories are seen in patients responding to other therapies such as rituximab and corticosteroids [1]. Therefore the main benefit in our patients was avoiding the side-effects of other therapies, such as splenectomy or prolonged high-dose steroid use, while allowing the natural history of acute ITP to run its course. This is even more relevant now with updated guidelines recommending short-courses of steroids (< 6 weeks), whereby there should be earlier movement towards second-line therapies if there has been an inadequate response.

Bleeding events in this study were infrequent in responding patients. Three patients with transaminitis resolved with drug cessation or dose reduction. Of note, these were patients all on an eltrombopag doses of > 100 mg and close monitoring of liver function should be undertaken in patients on higher doses. While the current dosing guideline of eltrombopag is the maximum dose of 75 mg, higher doses are used for other indications such as aplastic anemia. There were two cases of VTE treated with anticoagulation.

The strengths of our study are the multicenter design and long-term follow up. Limitations include the lack of a control arm with which to compare the rate of response to eltrombopag.

In conclusion, we found a high response rate to eltrombopag in a high-risk group of recently diagnosed steroid-refractory or dependent ITP patients, who have few other rapidly-acting treatment options. This may obviate the need for splenectomy and other more toxic therapies in a significant proportion of patients.

H.T. and A.G. were the principal investigators of the study, involved in study design, protocol development, data analysis and development of the manuscript. C.D. was involved in data analysis and development of the manuscript. J.R. was the statistician for the study and was involved in study design, development of the analysis plans, development of the manuscript. R.B., T.B., S.C., and S.H. were members of the steering committee and were involved in the data interpretation and manuscript development.

R.B. has been a speaker at Novartis and Sobi sponsored meetings, airfare and accommodation received, honorarium declined. Investigator in clinical trials sponsored by Novartis and Sanofi. JR is a former employee of Novartis AG and holds shares in the company. H.T., C.D., T.B., S.C., C.M., S.H. and A.G. have no relevant conflicts of interest to declare.