ESA are generally the first line treatment of anemia of IPSS low and int1 (lower) risk MDS. While endogenous EPO level and RBC transfusion requirement are well defined prognostic factors of response to ESA (Hellstrom, BJH 2003), many patients (pts) are treated before they require RBC transfusions and, in a GFM large cohort of lower risk MDS with anemia, baseline EPO level was below 500U/l in 88% of the cases (Kelaidi, Haematologica 2010). In non transfusion dependent (non TD) lower risk MDS, it is therefore important to find additional simple prognostic factors of response to ESA. In a joint study of the French (GFM) and Italian (FISM) MDS groups, we retrospectively assessed prognostic factors of response to ESA in non TD lower risk MDS. Particular emphasis was put on the prognostic value of marrow dysplastic features on ESA response, suggested in a previous work (Howe, Blood, 2004).
We are performing a cooperative (GFM and FISM) retrospective study in IPSS low and int 1 (lower risk) MDS diagnosed between 1994 and 2010. The present analysis focused on lower risk MDS with Hb level <10g/dl, non TD, and with baseline serum EPO <500, treated with ESA in France (from the GFM registry and from a previous cohort: Park, Blood 2008) and in Italy (Italian FISM registry). Pts received ESA 30,000 to 60,000IU/l /week(for EPO) or 150 to 300ug/week for Darbepoetin (DAR) during at least 12 weeks, with or without G-CSF. Response to ESA was evaluated after 12 weeks of treatment, using IWG 2006 criteria (responses before 2006 were reassessed). Pts receiving ESA combined with other drugs than G-CSF or after other drug treatment were excluded. The following baseline parameters were analyzed for their prognostic value on response to ESA: Hb, MCV, ANC and platelet count, serum ferritin (SF), serum EPO, WHO 2008 classification (earlier diagnoses were reclassified), IPSS, and precise quantification of marrow blasts, ring cells and dysplastic features on marrow aspirates (harmonized between 6 GFM and Italian morphologists for each pt according to WHO 2008 criteria).
For this work, 503 pts with adequate data have been identified. Results are presented in 228 completely reviewed cases. For morphology, there was an inter observer agreement in 85% of cases, discrepancies being resolved by taking mean values. Median values at onset of ESA in those pts were: age 75 (range 21–96), Hb 9.1g/dl (range 4,9–10), MCV 98u3 (85–127), ANC <1.8G/l in 33% pts, <0.8G/l in 8% pts, platelets <100G/l in 20% pts, SF 201ng/ml (range 20–1670), sEPO 29.5 IU/l (range 2–458), marrow blasts 1% (range 0–9). Dysplasia in > 10% marrow cells was present: for dyserythropoiesis (DE) in 188 (82%), dysgranulopoiesis (DG) in 129 (57%) for dysmegacaryopoiesis (DM) in 96 (43%) pts. WHO 2008 classification was: RA 36%/ RARS 12%/ RAEB-1 9%/ RCMD 36%/ RCUD (other than RA) 1%/ MDS 5q-5%/ unclassifiable 1%. IPSS was low in 49%, int-1 in 44%, NA in 7% pts. IWG 2006 response rate was 60,6%. In univariate analysis, Hb<9g/dl (p=0.002), DG (52% vs 71%, p=0.007), WHO (RA 70%, RCMD 58%, RARS 68%, RAEB-1 30%, MDS 5q– 30%, overall p=0.03), serum EPO >150 U/l (p=0.001) were significantly associated with lower response rate. Response was 61% and 58% in pts with and without DE (p=0.8), and 55% and 65% in pts with and without DM (p=0.1), and 58% in RCMD versus 69% in RA, RARS (p=0.08). DG was inversely correlated with ANC (p=0.003), was seen in 45% IPSS low and 69% int 1 (P=0.003), was correlated with WHO (RA 35%, RARS 48%, RAEB-1 60%, RCMD 85%, MDS 5q– 33%, p<0.0001), was more frequent in pts with DM (73% vs 44%, p<0.0001) but was independent of Hb level and serum EPO. In multivariate analysis absence of DG, Hb>9g/dl and EPO <150UI/l, were the only 3 factors associated with better response to ESA (p=0.01, p=0.006, p=0.004, respectively).
In this population of lower risk MDS with non transfusion dependent anemia, significant dysgranulopoiesis (DG) was, in addition to Hb<9g/dl, and EPO>150 UI/l, associated with lower response rate to ESA in multivariate analysis. DG was a stronger predictive factor of non response to ESA than the WHO classification and, contrary to the findings of Howe (Blood 2004), multilineage dysplasia had only borderline prognostic value for response to ESA. We are analyzing more patients in this retrospective cooperative study, also studying other potential prognostic parameters of response to ESA including somatic gene mutations.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.