Retrospective analysis of effects of transfusion status on response to decitabine and survival in patients with myelodysplastic syndromes.

Abstract

6617 Background: The DNA-hypomethylating agent decitabine is used to treat myelodysplastic syndromes (MDS). Methods: This retrospective analysis pooled a subset of patients (pts) with MDS from 2 decitabine trials: 3-day (phase III) and 5-day (phase II). Excluded pts had received an erythropoiesis-stimulating agent. Pts were stratified as RBC or PLT transfusion dependent (TD) or independent (TI) at baseline (BL) and on study. Response rate (RR, CR+PR+mCR) to decitabine (Fisher exact test) and overall survival (OS) were compared (log-rank test). Results: Pt characteristics, transfusion status, RR, OS, and time to TI are in the table. 23/62 (37%) RBC TD pts at BL became TI on study, and 10/18 (56%) PLT TD pts at BL became TI on study. RBC TI pts at BL and on study (24/34, 70%) had the highest RR (54%, p<.0001) and longest OS (median 19.8 mo; p=.0032).RBC TI pts, regardless of BL status, had significantly longer survival probability (HR 0.33; p=.0122). RBC or PLT TD pts at BL who did not achieve TI on study had the lowest RR and OS. Conclusions: In this retrospective analysis, MDS patients treated with decitabine who were TI at BL or achieved TI on study appeared to have better clinical response and better survival compared with patients who were TD at BL and did not achieve TI on study. Transfusion status N Age, mean, y IPSS high risk, n (%) ORR (CR+PR+mCR)n (%) CRn (%) PRn (%) mCRn (%) Overall survival, median (95% CI), mo Time from TD to TI, median (95% CI), mo Red blood cells BL TD/ on-study TD 39 71 11 (28) 1 (3) 0 1 (3) 0 9.4 (5.8, 13.0) NE BL TD/ on-study TI 23 69 3 (13) 9 (39) 8 (35) 1 (4) 0 19.2 (14.7, 21.6) 2.3 (0.99, 2.6) BL TI/ on-study TD 10 68 2 (20) 2 (20) 0 0 2 (20) 6.8 (0.99, NE) NE BL TI/ on-study TI 24 69 5 (21) 13 (54) 9 (38) 3 (13) 1 (4) 19.8 (12.9, NE) NE P <.0001* <.0001* .2994* .0156* .0032† <.0001† Platelets BL TD/ on-study TD 8 70 3 (38) 0 0 0 0 2.1 (0.9, 6.7) NE BL TD/ on-study TI 10 72 2 (20) 3 (30) 2 (20) 1 (1) 0 14.2 (5.0, 22.3) 2.9 (0.6, 15.5) BL TI/ on-study TD 19 69 4 (21) 1 (5) 0 0 1 (5) 3.1 (1.6, 12.6) NE BL TI/ on-study TI 59 69 12 (20) 21 (36) 15 (25) 4 (7) 2 (3) 19.2 (15.7, 25.7) NE P .0120* .0234* .5038* 1.000* <.0001† <.0001† Abbreviations: CI, confidence interval; NE, not evaluable. * Fisher exact test. † Log-rank test.