Background and objective. In a previous study we showed encouraging outcome in severe aplastic anemia (SAA) patients treated with anti-lymphocyte globulin (ALG), cyclosporin (CyA) and G-CSF 5mg/Kg/day (

Blood
2000
;
95
:
1931
–4.
). However failure to respond, delayed responses, partial responses, relapses and early deaths remain signifcant problems. The aim of the present study was to test whether an increased dose of G-CSF (10 mg/Kg/day) would reduce these complications.

Design and methods. This is a multicenter prospective trial in 77 SAA patients treated with horse ALG (15 mg/kg/day day1–5) and CyA (5 mg/kg/day day 1–180). Patients were randomized to receive G-CSF 5 mg/Kg/day (n=38, group A) or 10 mg/Kg/day (n=39, group B) from day +1 to day +30. All patients then received G-CSF 5 mg/Kg/day from day +31 to day +90. Primary end point was response at day +120. Secondary end points were early deaths , blood counts at day +120, and survival.

Results. At day +120 responses were classified as absent, partial, complete in 12, 22, 4 patients in group A and in 23, 7, 9 patients in group B (p=0.001). At last follow up these figures were respectively 9,12,17 vs 19,2,18 (p=0.004). Thirteen patients (5 in group A and 8 in group B) died before day 120 (p=0.3). Median peripheral blood counts an day 120 were comparable in the two groups: Hb 10.5 vs 9.5 gr% (p=0.6), Neutrophils 2.4 vs 1.9x10^9/l (p=0.4) and platelets 42 vs 36 x10^9/l (p=0.3). The actuarial survival at 4 years is 72% in group A vs 67% in group B (p=0.3).

An additional finding of this study is a strong age effect, with an actuarial 4 year survival of 81% in patients aged 0–20 , 80% in patients aged 21–40 and 34% in patients over 40 (p=0.0002). This correlated with the inability of older patients to increase their white blood cell (WBC) counts above 5x10^9/L, during G-CSF treatment. In a multivariate COX analysis patient age and highest WBC counts during G-CSF, were both significant predictors of survival.

Interpretation and conclusions. Increasing the dose of G-CSF does not appear to reduce early deaths, does not improve peripheral blood counts nor survival, and may reduce the response rate in patients with SAA receiving ALG and CyA. Older age and failure to improve WBC are negative predictors of survival.

Author notes

Corresponding author

Sign in via your Institution