Characteristics of romiplostim-treated MDS patients with hematologic improvement in platelets (HI-P)

Background: Thrombocytopenia occurs in 40%-65% of MDS patients. Romiplostim increases platelet counts and decreases bleeding in MDS. We examined romiplostim-treated MDS patients achieving HI-P per IWG 2006 criteria for 8 weeks (baseline >20×109/L: increase ≥30×109/L; baseline <20×109/L: increase to ≥20×109/L and by ≥100%). Methods: Thrombocytopenic IPSS low/int-1 risk MDS patients received romiplostim at fixed doses (300-1500 μg) in two phase 2 trials (NCT00303472, NCT00614523) and in doses (250-1500 μg) adjusted for platelet counts in an open-label extension (NCT00472290), weekly or every other week, and generally subcutaneously (intravenously in one study arm). A post hoc analysis was performed on the 34 of 60 romiplostim-treated patients achieving HI-P. Results: 53% of the 34 patients achieving HI-P were women, 82% were white, and median (Q1, Q3) age was 70 (65, 78) years. WHO category at diagnosis was RCMD (41%), RA (32%), unclassified (15%), RAEB-1 (6%), RARS (3%), and del(5q) (3%). IPSS status was low (29%), int-1 (65%), or unknown (6%). Median (Q1, Q3) baseline platelet count was 31.5×109/L (20.0, 45.0); 21% of patients had baseline platelet counts <20×109/L. The median (Q1, Q3) longest continuous HI-P response was 28 (14, 56) weeks. At last platelet count, 32/34 patients met HI-P criteria. Median platelet counts were >50×109/L from week 1 on. Median (Q1, Q3) number of doses was 42.5 (21, 125), median (Q1, Q3) exposure duration (first to last dose) was 43.5 (23, 160) weeks, and median (Q1, Q3) average weekly dose was 750 (648, 950) μg. Most (85%) patients did not have grade ≥3 bleeding events; 3 (9%) had one grade ≥3 bleeding event and 2 (6%) had two grade ≥3 bleeding events (at study weeks 10, 39, 40, 45, 78, 142, and 142). AML occurred in 2 patients (after 44 and 46 weeks of romiplostim; baseline WHO categories RCMD and RAEB-1, respectively) for a rate of 3.6/100 patient-years. No deaths were reported. In a logistic regression model of patients with ≥10 weeks of treatment in the extension (N=48), only higher baseline platelet count (≥20×109/L) was associated with developing HI-P (odds ratio=4.2, 95%CI:1.1-15.8, unadjusted P=0.03); also tested were age >75 years, sex, race, ECOG, and WHO category. Conclusion: HI-P observed in MDS patients receiving romiplostim was associated with higher baseline platelet count. (Figure Presented).