OBJECTIVETo compare the efficacy and safety of four different regimens for pediatric severe aplastic anemia (SAA) with immuno-suppressive therapy (IST) with or without combined human granulocyte colony-stimulating factor (G-CSF).

METHODThe authors retrospectively analyzed 105 children with SAA treated with IST with or without G-CSF in the hospital from February 2000 to September 2010. Regimen A, without G-CSF in the whole treatment, was used to treat Group A patients, n = 27; Regimen B, G-CSF, was initiated in Group B, n = 24, before the IST until hematologic recovery; Regimen C, G-CSF, was used together with the IST for Group C patients, n = 24, until hematologic recovery; Regimen D,G-CSF was used for Group D, n = 30, after the end of IST until hematologic recovery. The response rate, relapse rate, mortality, infection rate, infection-related death rate, risk of evolving into MDS/AML, survival rate, factors affecting the time of event-free survival and so on.

RESULT(1) The response (CR+PR) rates 4, 6, 12 and 24 months after IST of the whole series of 105 SAA children were 50.5% (7.6%+42.9%) , 60.0% (21.9%+38.1%) , 67.6% (38.1%+29.5%) and 69.5% (40.0%+29.5%) respectively. The 2-year survival rate was 90.5%; the follow-up of the patients for 13 years showed that the whole survival rate was 87.6%. (2) The differences of the response rates 4, 6, 12 and 24 months after IST of the 4 groups were not significant (P > 0.05). (3) No significant differences were found in the mortalities 4, 6, 12 and 24 months among the 4 groups (P > 0.05). (4) Of the 105 patients, 4 children had relapsed disease in the period of time from 6 to 24 months after IST. All the four patients belonged to the groups with G-CSF. (5) The use of G-CSF could not decrease the infection period before IST (day) (P = 0.273), and it had no impact on the infection rate after IST (P = 0.066). It did not reduce the rates of septicemia and infectious shock. And to the infection-related death rate no significant conclusion can be made. (6) Follow up of the patients for 13 years, showed that 2 had the evolution to MDS/AML in the 105 patients and the two children belonged to the groups with G-CSF. (7) Kaplan-meier curve analysis did not show any differences in the survival rates of the four groups. (8) Cox regression analysis showed that the use of G-CSF had no benefit to the patients' long term survival. While the age of diagnosis and the infection history before IST were significantly related to the patients' long term survival.

CONCLUSIONThe use of G-CSF did not contribute to the early response and could not reduce the infection rate, infection-related death rate and the patients' long term survival. There were no significant differences in the survival rates of the four groups. Attention should be paid to the risk of the evolution to MDS/AML.