Prediction of Response to Recombinant Erythropoietin Plus Granulocyte-Colony Stimulating Factor Following a Single Subcutaneous Bolus in Patients with Myelodysplastic Syndromes; a Randomised Placebo Controlled Study.

Recombinant Erythropoietin (+/− G-CSF) is an effective therapy for the anaemia of selected patients with MDS. Validated response prediction models are available, but response rates are only 60% in the “high” predicted response group. Furthermore, half of the total cost of one year’s therapy for a cohort of patients selected for intermediate / high predicted response, is incurred within the initial 12-week therapeutic trial (Cassadeval et al, Blood 2004,104;321). Our hypothesis was that the erythroid response to a single bolus of EPO + G-CSF (Part 1) may predict for sustained response to a therapeutic trial (Part 2). 21 MDS patients (<10% blasts) were randomised in Part 1 to receive either a single s.c. bolus of EPO 18 000 units (NeoRecormon) plus G-CSF (Lenograstim) 263 mcg (n=10), or two vials s.c. placebo (n=11). Serum EPO, haemoglobin concentration and reticulocytes (Sysmex SE9000) were assayed daily from Days 1–8. 20 patients proceeded to Part 2 and received an 8 week therapeutic trial of s.c. EPO 9000 units thrice weekly (tiw), weeks 1–4, escalating to 18 000 units tiw weeks 5–8 if no response, plus titrated s.c. G-CSF tiw. Responders were changed to once weekly (qw) EPO dosing from weeks 12–20 at the total weekly responding dose. 6 patients had erythroid response by study response criteria and 7 by IWG criteria (2HI-E major, 5 HI-E minor). 4/7 RARS patients responded. Incremental change in absolute reticulocyte counts between Day 1 and Day 8 of Part 1 discriminated responders (median increment = 40x109/l, range 31–81, n=6), who received bolus EPO/G-CSF, from non-responders who also received bolus EPO/GSCF (median increment = 1.5x109/l, range −14 to 6, n=4) and from patients receiving placebo (median increment = 5x109/l, range −21 to 18, n=11)(ANOVA P=.002). An incremental increase of >30x109/l was 100% predictive of subsequent response. In patients with erythroid response in Part 2, haemoglobin concentration at qw EPO either did not change compared to tiw dosing (P>.05, n=5), or increased (P=.002, n=1). Serum ferritin, transferrin saturation, CHr (Bayer Advia) and serum transferrin receptor (TfR)concentrations were assayed weekly. Two patients became biochemically iron deficient during weeks 1–8, both of whom had baseline serum ferritin <100mg/l. No iron supplementation was given, and one patient still had an erythroid response. No clear evidence for functional iron deficiency was seen in patients with serum ferritin >100 mg/l. Serum non-transferrin bound iron concentration correlated closely with transferrin saturation both at baseline (n=21 patients), and on treatment (n=4 responders and 4 non-responders). In Part 2, neither ΔHb, nor ΔTfR at weeks 1 or 2 predicted response. No baseline erythroid parameters differed between responders and non-responders. New observations: 1. Absolute reticulocyte increment at Day 8 post s.c. bolus EPO/G-CSF predicts for therapeutic response in this small study, 2. Once weekly EPO is as effective as thrice weekly EPO in similar doses, 3. Functional iron deficiency may impair response in MDS patients with iron-limited erythropoiesis.