We recently reported on the risk for death from sepsis and myelodysplastic syndrome/acute myeloid leukemia (MDS/AML) in patients with severe congenital neutropenia (SCN) on long-term granulocyte-colony stimulating factor (G-CSF) enrolled in the Severe Chronic Neutropenia International Registry (SCNIR). (

Rosenberg et al,
Blood
2006
;
107
:
4628
-35
, and
Brit J Haematol
2008
;
140
:
210
-3
) We found that after ten years a subgroup of SCN patients receiving more than 8 mcg/kg/day of G-CSF who achieved a median absolute neutrophil count (ANC) response of less than 2.188 × 109/L had an estimated risk of dying from sepsis of 14% and a risk of developing MDS/AML of 40%. In order to understand the risk factors for death from sepsis, we reviewed the clinical diagnosis, hematological data, G-CSF treatment, and clinical course of 9 fatalities due to infections, excluding cases with evidence of MDS or AML or occurring after hematopoietic transplantation. The 5 male and 4 female patients’ ages ranges from 1 month to 18.7 years at the time of initiation of G-CSF. All were treated with G-CSF because of severe neutropenia (ANC less that 0.5 × 109/L) and a history of recurrent fevers and infections. The pre G-CSF median ANC for the 9 patients was 0.027 × 109/L and frequent events pre-G-CSF included fevers, severe mouth ulcers, otitis, sinusitis, pneumonia, cellulitis and abscess formation, but bacteremias were infrequent. When started on G-CSF, five patients were clearly poor responders, despite steady upward titration in the G-CSF dose to 22, 45, 65, 68 and 151 mcg/kg/day, respectively. In each of these patients, the ANCs on G-CSF treatment were frequently reported as less than 0.1 × 109/L and the responses were inconsistent and quite variable. The other 4 patients also had inconsistent responses at lower G-CSF doses and the patients’ compliance with treatment was uncertain. Two are known to have had a protracted “off treatment” periods, and the septic death of one patient occurred when off therapy. Deaths were attributed to sepsis (6 cases), pneumonia (2 cases), and meningitis (1 case). These clinical data suggest that death from sepsis in SCN is usually associated with a poor response to G-CSF as reflected by the requirement for dose escalation substantially above the threshold of 8 mcg/kg/day of G-CSF associated with an increased risk of MDS/AML. Inconsistent treatment and discontinuation of treatment also lead to neutropenia and the risk of sepsis. Although impaired microbicidal function may be a contributor to infections in these patients (
Donini, et al.
Blood
109
:
4716
,
2007
), the predominant problem is a poor ANC response; the biological basis for poor responses of some SCN patients to G-CSF is still unknown. Based on these data, we recommend careful consideration of a match related or unrelated donor hematopoietic stem cell transplant for all patients not achieving a consistent ANC >1.0 × 109 with G-CSF therapy.

Disclosures: Boxer:Amgen: Equity Ownership. Dale:Amgen: Consultancy; NIH: Research Funding; Merck: Research Funding; Genzyme: Research Funding; Cellerant: Consultancy; Telik: Consultancy; Schering-Plough: Consultancy; Caremark: Consultancy.

Author notes

Corresponding author

Sign in via your Institution