Thrombopoietin Receptor Agonist (TPO-RA) Switch in Adult Primary Immune Thrombocytopenia (ITP) Patients: A Retrospective Collaborative Survey from 8 Italian Hematology Centers

Introduction. Availability of the 2 TPO-RAs Romiplostim (Rom) and Eltrombopag (El) offers an effective treatment option for primary ITP patients (pts). However, some pts are either not responsive or lose response - i.e. desired platelet (plt) count achieved but not sustained over time, or experience wide fluctuations in plt count with either TPO-RA. Adverse events may cause treatment discontinuation. Finally, ptÕs preference may be an important issue considering the different route and timing of administration of the two agents. Availability of two TPO-RAs for clinical use, with different molecular structure and site of binding within the TPO receptor, makes it appealing to try switching with the aim of overcoming treatment limitations of either agent. The present survey offers insight into outcome of TPO-RA switching in a group of ITP pts treated at 8 Centers representative of the Italian territory.

Patients. Charts of ITP pts on treatment who underwent TPO-RA switch were retrospectively reviewed.

Results. Between Jan 2009 and Feb 2015, 57 of 249 pts on TPO-RA (22,9%) underwent switch: El ˆ Rom 26/57 (45.6%), Rom ˆEl 31/57 (54.4%). Median age at 1^st TPO-RA administration was 55 yrs (range 16-81); M/F = 23/34. Median disease duration prior to 1^st TPO: 58 mos (range 2-648). Median lines of previous therapy 3 (range 1-6; splenectomy: 23/57, 40.4%). Overall 42/57 pts (73,7%) had received maximum product dose as per prescribing information prior to switch. Table 1 summarizes reasons for TPO-RA switch and outcome. Overall, 32/57 pts (56.1%) achieved, regained or maintained a response upon switching. The majority of pts (39/57, 68.5%) were switched for efficacy issues, i.e. failure to respond to 1^st TPO (27 pts) or response loss (12 pts); among these 39 pts, 48.7% responded to the 2^nd TPO-RA. One pt lost response to Rom because of development of neutralizing antibodies; response was regained upon switching to El.

In this subgroup of pts, disease duration and lines of previous therapy (but not splenectomy status) seem to have an impact on response to switching. Each one month increase in disease duration determines a 0.7% decrease in the odds of achieving a response (WaldÕs test p=0.065). More than 2 lines of therapy determine a 72% decrease in the odds of achieving a response (WaldÕs test p=0.077). Either TPO-RA switch sequence was equally effective in yielding response (FisherÕs exact test p=0.752) and age at 1^st TPO-RA had no impact on response. Of the 18 pts switched for reasons other than efficacy, 13 (72.2%) maintained a response on the 2^nd TPO-RA: 5/6 switched for plt count instability (counts stabilized = 2/5 responding pts), 4/7 switched for ptÕs preference, 4/5 switched for side effects. Four pts (1 plt count instability, 3 ptÕs preference) underwent Òdouble switchÓ (i.e. Rom ˆ El ˆ Rom): re-exposure to Rom was not associated with response loss.

Discussion. Switching enables approximately 56% of pts to achieve, regain or maintain a plt response; switching for inefficacy yields lower response rates (48.7%) compared to switching for reasons other than efficacy (72.2%) Plt counts fluctuation stabilized in 40% of pts. Re-exposure to R in the 4 pts who underwent Òdouble switchÓ was not associated with response loss, confirming absence of tachyphylaxis with this TPO-RA. Our results are in line with those reported by Khellaf (Haematologica 2013) and Gonzales-Porras (BJH 2014): TPO-RA switch can be a safe and appealing treatment option for ITP pts who experience suboptimal results with either agent.