Melvin H. Freedman, M.D., Division of Hematology-Oncology, Department of Pediatrics, University of Toronto, Faculty of Medicine, Toronto, Ontario, Canada, 1997

When a new product with huge clinical potential explodes on the scene, the hope is that the benefits far outweigh the risks in long-term administration. After 10 years of clinical use, granulocyte colony-stimulating factor (G-CSF) has lived up to that promise so far. In the context of severe chronic neutropenia, more than 90% of patients have reaped big benefits in terms of improved quality of life and less infection, inflammation, antibiotic use, and hospitalization as well as oropharyngeal ulcers. With long-term use, toxic and adverse events have been catalogued but in general are not clinically troublesome and, aside from occasional adjustment of scheduling and dosing, seldom necessitate stopping therapy. Currently the topic of intense focus is the phenomenon of malignant myeloid transformation in patients with congenital neutropenia who are receiving G-CSF. Data from the Severe Chronic Neutropenia International Registry have identified 23 of 249 patients with congenital neutropenia who have developed myelodysplasia or acute myelogenous leukemia (MDS/AML) giving a crude rate of about 9% with an average follow-up of 4.5 years. No cases of MDS/AML have occurred in 257 patients with cyclic or idiopathic neutropenia. A critical analysis of the incidence of transformation year by year shoed a fairly uniform hazard rate of less than 2% per year, and the risk of MDS/AML after 5 to 6 years of therapy did not appear to be greater than during the first year of therapy. The transformation risk in the congenital cohort must also be viewed in the context of published reports of spontaneous myelodysplasia or acute myelogenous leukemia occurring in these patients in the pre-G-CSF era. Thus, the role of G-CSF in malignant conversion is still not clear and requires long-term vigilance and research. G-CSF is still deemed specific therapy for severe chronic neutropenia with a high margin of safety and should be the initial treatment for this family of disorders.

Brief preclinical history of granulocyte colony-stimulating factor

With assays similar to those used to characterize murine granulocyte colony-stimulating factor (G-CSF), an activity was isolated in 1985 from a medium conditioned by the growth of the 5637 bladder carcinoma cell line and named pluipoietin by Welte et. Al. because of its ability to stimulate the proliferation and development of multipotent myeloid progenitor cells in vitro. When the activity was cloned and expressed as a recombinant protein in 1986, it was demonstrated that the pluripoietin property was lost and that the activity was apparently restricted to granulopoiesis. This resulted in the molecule being renamed G-CSF.

This major discovery triggered worldwide interest in scientific and medical communities and prompted preclinical studies in which the in vivo effects of G-CSF were defined. The potential for clinical application was confirmed in studies showing that G-CSF increased peripheral blood neutrophil numbers in vivo when injected in rats, mice, hamsters, dogs, and humans and that accelerated recovery of neutrophils occurred after cyclophosphamide-induced myelosuppression in mice and monkeys.

Significantly, the exposure of animals to high doses of G-CSF for protracted periods did not appear to have deleterious effects. For example, high constitutive expression of G-CSF, attained after bone marrow transplantation with hematopoietic cells that had been infected with a retroviral expression vector containing the gene encoding G-CSF, resulted in sustained neutrophilia and infiltration of tissues, especially lung and liver, with neutrophils. Similar infiltration was noted in mice that had been repeatedly injected with G-CSF. In both studies, however, this did not result in morbidity, and no obvious pathologic changes were seen.

Efficacy versus safety in phase I/II and III trials

Severe chronic neutropenia (SCN) was targeted early in 1987 for clinical trials. The designation SCN refers to a heterogeneous group of hematologic disorders associated with a selective decrease in peripheral blood neutrophil numbers to levels that predispose to chronic oropharyngeal inflammation, recurrent fevers, and severe infections. For ease of classification, SCN is divided into three main syndromes; idiopathic neutropenia, with onset an any age and for which no etiology is evident; cyclic neutropenia, with regular predictable oscillations in numbers of neutrophils and other cellular lineages; and congenital neutropenia, with early onset of abnormal hematological finds, often at birth. Within the congenital category, Postman’s syndrome, or congenital agranulocytosis, is a subtype inherited in an autosomal recessive manner. Other genetic forms of congenital neutropenia are Shwachman syndrome and glycogen storage disease type 1b.

Two open-label, slight-center phase I/II trials in patients with SCN showed that G-CSF doses of 3 to 60 m g/kg/d effectively increased blood neutrophils and decreased infection and inflammation in all cases and that long-term maintenance treatment for a minimum of 100 days and up to 13 months sustained the responses. In addition, six patients with cyclic neutropenia were treated in a third study with G-CSF, 3 to 10 m g/kg/d, for 2 months or more and demonstrated that the oscillations of neutrophil counts persisted on therapy but the cycle length shifted from 21 to 14 days with marked abbreviation of the duration of severe neutropenia.

Although these phase I/II trials were designed for short-term and intermediate-term G-CSF administration, clinical safety was carefully monitored. In two of the trials, bone pain was experienced after intravenous administration of G-CSF but abated with subcutaneous administration. In two of five patients in the first series and all six patients in one of the other trials, a moderate increase in the size of the spleen was noted but did not cause a clinical problem. A consistent biochemical finding was an elevated leukocyte alkaline phosphatase level while patients were receiving G-CSF therapy. Since leukocyte alkaline phosphatase is a marker of postmiototic granulocytes, the increased activity likely reflected the growth-promoting effect of G-CSF on granulopoiesis and is therefore not in itself harmful. Of note, one patient experienced a cutaneous necrotizing vasculities (leukocytoclastic vasculitis) on both legs at day 12 of G-CSF therapy, which abated on stopping the cytokine, reappeared on restarting it, and was then controlled by carefully titrating the dose in accordance with the clinical picture.

Because of the beneficial neutrophil responses from G-CSF in SCN in these phase I/II studies and corroborative case reports and additional trials, a water-shed phase III randomized, controlled trial was initiated in the United States in late 1988; the convincing results showing clear-cut efficacy of G-CSF treatment for SCN were published in 1993. The trial enrolled 123 patients (60 with the congenital form, 21 with the cyclic form, and 42 with the idiopathic form) of which 120 had evaluable responses. With a complete response over the 4 month treatment period and beyond defined as the maintenance of median neutrophil numbers greater than 1.5 x 109 cells/L, 90% of patients shoed a complete response to G-CSF therapy, and infection-related events and antibiotic use were significantly decreased. Safety data were analyzed and events related to the treatment were generally mild and consisted of headache, general musculoskeletal pain, transient bone pain, and rash. None of these required the discontinuation of G-CSF.

On the basis of the phase I/II studies showing that chronic administration of G-CSF increased spelling size presumably due to extramedullary hematopoiesis, splenomegaly was monitored carefully in the phase III trial by physical examinations or by magnetic resonance imaging or CT. Before therapy, 18 patients had an enlarged spleen on palpation, ranging from 1 to 4 cm below the left costal margin. With therapy, palpable splenomegaly was reported in 29b patients, ranging from 1 to 5 cm. In five of nine participating centers, splenic volume was assessed by CT or magnetic resonance imaging. The median percent increase in splenic volume among 59 patients evaluated was 34% (range, 2% to 148%). The increase in the spleen size was generally asymptomatic, and there was no evidence of infarction, hemorrhage, or predictable effects of the splenic enlargement on other blood cell counts during the 5 month treatment period. No splenectomies were required.

Excluding patients with a single platelet count below 100 X 109 cells/L, thrombocytopenia (defined as ³ two counts less than 100 x 109 cells/L) was noted in seven patients. Four had documented prestudy platelet counts lower than 100 x 109 cells/L. Two cases were mild (>50 x 109 cells/L), two were moderate (25 to 50 x 109 cells/L), and three more were severe (<25 x 109 cells/L). Two of the three had Shwachman syndrome and histories of thrombocytopenia prior to therapy. Although these patients required discontinuation of G-CSF for a few days, all subsequently returned to daily therapy.

Thus, the vast majority of patients in this phase III trial benefited substantially from G-CSF therapy, with minimal adverse or toxic effects, and almost all of the originally treated patients entered a long term G-CSF maintenance program. Reduction in the number of fevers, infections, and inflammation translated into a well-documented improvement in the quality of life for these patients.

Reports on safety from related trials

Because G-CSF is now used mostly to accelerate marrow recovery after chemotherapy for solid tumors and hematologic malignancies, the cytokine is administered only on a short-term basis in this setting and its long-term safety cannot be assessed. The only current reports that address long-term use of G-CSF fall into two categories: acquired and constitutional (Fanconi’s) aplastic anemia, and SCN.

In a large multicenter G-CSF trial for acquired aplastic anemia, 27 patients received cytokine for 2 to 12+ months in doses from 100 to 400 m g/d subcutaneously or 250 to 1500 m g/d intravenously. Of the 27 patients, 26 showed neutrophil response, 10 had increased hemoglobin levels, and 3 of these 10 thrombocytopenia improved as well. These multilineage responses were not associated with clinically significant morbidity or adverse effects except for mild bone pain at initiation of therapy, and the over improved clinical status of the responders was impressive. Two patients in whom neutropenia improved with G-CSF therapy died of splenic rupture and pulmonary hemorrhage, respectively, but these complications could not be related directly to the treatment. This report substantiated a previous G-CSF study, and was confirmed in a subsequent G-CSF (plus erythopoietin) trial for aplastic anemia which also demonstrated recovery of one or more hematopoietic lineages in some patients treated for up to 47 months with apparent safety. Aside from elevation of serum leukocyte alkaline phosphatase in 53% of patients that had no clinical consequence, no toxic effect, allergic reaction, or antibody formation was observed, and there was no progression of disease in any patient.

It is well-known that between 13% and 20% of patients with acquired aplastic anemia who respond to immunosuppressive treatment transforms into a clonal myelopathy with myelodysplasia or acute myelogenous leukemia(MDS/AML). With this background of malignant propensity in aplastic anemia, it is not surprising that some aplastic patients successfully treated with G-CSF or other cytokines will also manifest MDS/AML with time. However, the role of G-CSF in the transformation event is obviously not clear yet and requires careful vigilance.

A very important clinical study examined the effects of prolonged administration of G-CSF in 12 patients with neutropenia caused by Fanconi’s anemia. The patients were treated with varying subcutaneous dosages daily or every other day for 40 weeks. By week 8 of the study, all patients had an increase in absolute neutrophil numbers and four had an increase in platelet counts. In addition, four patients who were not being transfused had a significant increment in hemoglobin levels and a fifth patient no longer required transfusion. Concurrent with the impressive improvements in hematology patients who finished 40 weeks of G-CSF treatment shoed increases in the percentage of marrow and peripheral blood CD34+ cells.

Genomic instability and a marked predisposition to leukemia and cancer are features of Fanconi’s anemia and therefore the wisdom of using a growth-promoting cytokine on a long-term basis for this disorder is a central issue. In this study one patient had a marrow clonal cytogenetic abnormality (48 XXY, +14) without MDS/AML at week 40 of G-CSF treatment. Therapy was stopped and within 3 months the +X, +14 clone disappeared, but monosomy 7 appeared in 11% of metaphases and increased over the ensuing months, prompting a bone marrow transplantation. Because monosomy 7 manifested and progressed after G-CSF was stopped, it seems highly unlikely that G-CSF was involved in the transformation event in this case. Also, the appearance of the +X, +14 clone while receiving G-CSF treatment and its disappearance on discontinuing the cytokine must be put into context of Fanconi’s anemia, in which clones, seemingly unstable, are known to manifest and disappear spontaneously.

Thus, this pilot study showed overall that G-CSF is probably a safe and effective form of treatment for Fanconi’s anemia. One patient had a low-grade fever that resolved with dosage modification, and no other untoward clinical effects were noted in any patients.

Data from the Severe Chronic Neutropenia International Registry

The Severe Chronic Neutropenia International Registry (SCNIR) was established in 1994 to monitor the clinical course, treatment, and disease outcomes in patients with SCN. The Registry is a valuable resource because of large numbers of patients entered into its worldwide database. Patient data are submitted internationally to the coordinating centers at the University of Washington, Seattle, USA, and the Medizinische Hochschule, Hannover, Germany. Currently, short-term and long-term information dating back to 1987 on a total of 506 patients is available for analysis. Patients enrolled in the SCNIR were reported recently in three detailed publications that addressed long term safety issues of G-CSF for SCN.

In the SCNIR 1995 annual report (on file at Clinical Safety, AMGEN Boulder, 3200 Walnut Street, Boulder, CO 80301), the consistent sustained hematologic response in patients treated with G-CSF for more than 6 years was confirmed. With therapy, neutrophil counts rose in more than 90% of SCN patients and were maintained at a plateau for protracted periods resulting in vast clinical benefits. There have been no cases of marrow or hematopoietic lineage "exhaustion" or depletion with G-CSF therapy, although 6% or less of patients had thrombocytopenia (<50 X 109 platelets/L) during treatment but most of them had a preexisting history of low platelets. Thus, thrombocytopenia remains an infrequent problem in SCN during G-CSF therapy.

Splenomegaly

As noted in the phase I/II and III trials quoted herein, palpable splenic enlargement occurs in patients with G-CSF therapy but it can exist prior to treatment as well. Patients with congenital neutropenia experienced the highest incidence of palpable splenomegaly at baseline (24%) with a median spleen measurement of 2cm below the left costal margin. During the first year of G-CSF administration, the incidence increased to 44% and thereafter the rate of occurrence of splenomegaly remained stable for this cohort. Baseline splenomegaly was noted in about 9% and 14% of cyclic and idiopathic patients, respectively, but the prevalence and degree of palpable splenomegaly remained less for these two groups compared to the congenital neutropenia patients during G-CSF therapy. One patient with neutropenia associated with glycogen storage disease type 1b developed splenomegaly and underwent splenectomy after 27 months of G-CSF. The histology showed extramedullary hematopoiesis. Whereas an increase in spleen size may be an effect of G-CSF therapy, the change does appear to be progressive or clinically problematic in most cases.

Osteopenia or Osteoporosis

By 1995, 26 patients (25 congenital and 1 idiopathic type) were enrolled in the Registry with osteopenia or osteoporosis. In a pilot study reported to the SCNIR by Karl Welte, 30 congenital neutropenia patients were examined for bone mineral content by quantitative CT. Fifteen of these 30 patient had bone mineral content below two standard deviations of the normal age-matched values. Two patients suffered from vertebral deformation and compression fractures. Preliminary results of bone biopsies revealed increased numbers of activated osteoclasts; cells derived from the myeloid progenitor cells. Although it is likely that the bone loss was caused by the patients’ underlying disease, it could not be excluded that treatment with G-CSF activates osteoclasts. Six of the 15 patients demonstrated bone loss prior to G-CSF treatment and did not experience significant change in bone mineral content during G-CSF treatment. This observation plus the fact that most patients receiving G-CSF for treatment of SCN have no clinical problems with fractures, decreased stature, or related problems suggests that G-CSF is not a direct cause of significant osteopenia or osteoporosis.

Vasculitis

Using combined data from the phase I/II and III clinical trials, a 3% incidence of cutaneous vasculitis was catalogued by the Registry in patients treated with G-CSF. These cased occurred in patients with idiopathic, cyclic, and congenital neutropenia. The lesions were usually limited to the skin and over half of these cases were biopsy-proven leukocytoclastic vasculitis. The vasculitis generally developed simultaneously with an increase in neutrophil numbers and abated when the neutrophils decreased. Most patients were able to continue G-CSF therapy at the same or a reduced dose. The mechanism of vasculitis in these patients is unknown.

Growth and development

Severe chronic neutropenia is not known to have a direct relationship on growth and development, although growth may be retarded by chronic infection or inflammation associated with neutropenia particularly in the most severely affected patients. Because the Registry involves children, many in their peak years for development of height and other developmental characteristics, a careful review of growth and development is on file. Analysis in 1995 showed that the percentage of children from the United States with SCN who are over the 50th percentile of weight for height increased with the duration of G-CSF treatment. These data also show that for children starting treatment before 3 years of age, the percentage with height over the 50th percentile appears to increase with duration of G-CSF treatment compared to those starting treatment after 3 years of age. This is the first evidence that early treatment may benefit early physical development in children with congenital neutropenia who tend to have retarded linear growth.

Pregnancy and Fertility

The SCNIR has compiled information on 11 pregnancies occurring in 9 patients with SCN treated with C-CSF, all of whom had either cyclic or idiopathic neutropenia. Six of these 11 pregnancies resulted in either a normal birth of a child with cyclic neutropenia, which is known to be inherited in an autosomal dominant fashion. In addition, the Registry has data on one male patient (with cyclic neutropenia) who fathered a child with cyclic neutropenia. In three other cases, congenital anomalies were detected either during gestation or at birth, and one of these pregnancies, plus two others without known congenital anomalies resulted in elective terminations. Of the two live births with congenital abnormalities, one had bilateral hydronephrosis and the other a cardiac defect. Both of these developmental abnormalities are relatively common; at present there is no known association with G-CSF treatment. Thus, the overall effects of G-CSF in pregnancy are not yet known because the data from these few case reports are inconclusive.

Malignant myeloid transformation in severe chronic neutropenia patients receiving granulocyte colony-stimulating factor

Currently, the topic of concern is the phenomenon of malignant myeloid transformation in SCN patients receiving G-CSF therapy. Table 1 shows the most up-to-date Registry information on MDS/AML in the database. Of 249 patients classified as congenital neutropenia(including Kostmann’s syndrome), 23 have developed MDS/AML yielding an overall incidence or crude rate of about 9% with an average follow up of 4.5 years. It is noteworthy that no cases of MDS/AML occurred in the subgroup of congenital neutropenia patients with glycogen storage disease type 1b, nor with cyclic or idiopathic neutropenia. However, when all patients are tallied (n=506), the crude rate is 4.5%

Cases of patients with chronic neutropenia who developed

MDS/AML while receiving G-CSF entered into the Registry

Table 1

  n MDS/AML Incidence%
Congenital 249 0 0
Idiopathic 160 0 0
Total 506 23 4.5
       

A crucial issue is whether long-term administration of G-CSF for SCN increases the risk of transformation. Do the chances of the event escalate with time? To answer this the first 15 cases of MDS/AML were analyzed year by year of therapy starting from the time interval of 0 to 1 year and proceeding stepwise up to year 6 of G-CSF exposure. Beyond year 6 the data are inadequate for analysis and therefore only the first 15 cases of MDS/AML could be evaluated. Table 2 shows the number of patients taking G-CSF at the start of each time interval, the number of cases of MDS/AML occurring in that interval, and the percent or hazard rate determined by numerator over denominator with appropriate statistical adjustments to account for denominator changes. As shown, the malignant conversions and hazard rates are fairly uniform for each time span, less than 2% per year, except for an apparent increase in year 4 to 5. The meaning of this increase is not clear because the pattern was not reproduced in the cohort analyzed at the 5 to 6 year interval of therapy. Note also that the number of converters in each interval is small (for example, n=3 at 0-1 year versus n=5 at 4-5 years) and statistical differences cannot be confidently determined. Thus, at our current level of understanding in 1997, the hazard rate for MDS/AML at 5 to 6 years of G-CSF treatment for SCN does not appear to be greater than during the first year of therapy. However, as of May 1997, ongoing analysis of the risk of MDS/AML with respect to duration of therapy disclosed seven new cases of transformation in congenital neutropenia patients treated with G-CSF for 6 to 7 years. Thus the hazard rates for transformation appears uniform for the first years of treatment but escalate to 11% at year 7 of therapy. Patients were also analyzed for age at presentation, sex, and mean G-CSF dosage, and no relationship with any of these variables was apparent either.

Table 2

Conversion to MDS/AML in relationship to duration of G-CSF therapy for severe chronic neutropenia

Interval, y n at start MDS/AML Percent*
0-1 191 3 1.6
1-2 174 2 1.2
2-3 155 2 1.3
3-4 147 2 1.4
4-5 134 5 4.6
5-6 104 1 1.2

*Adjusted for patients not completing interval for reasons other than leukemia. AML-acute myelogenous leukemia, G-CSF - granulocyte colony-stimulating factor, MDS-myelodysplasia

Aberrant cellular genetic changes

Conversion to MDS/AML in the SCN patients was associated with one or more cellular genetic abnormalities that may be useful in identifying a subgroup of patients at high risk. Of 23 who had malignant myeloid transformation in the Registry series, 14 developed partial or complete loss of chromosome 7 (7q-or monosomy 7) in marrow cells; none of the patients who were tested prior to G-CSF therapy had monosomy 7. Activating ras oncogene mutations were discovered in 5 of 10 patients from the series of 23 after the transformation to MDS/AML but not before. Four of these also had monosomy 7. The mutated fragments were cloned and sequenced and showed GGT (glycine) to GAT (aspartic acid) substitutions at codon 12 in all of them (K12ASP in 2 patients and N12ASP in 3 patients). Marrow cells from five patients with malignant myeloid transformations also showed point mutations in the gene for G-CSF receptor, resulting in a truncated C-terminal cytoplasmic region of the receptor that is crucial for maturation signaling. Twenty patients without receptor mutations showed no evidence of progression to MDS/AML, but four additional patients have been identified with mutations but without MDS/AML and are currently being monitored closely.

Role of granulocyte colony-stimulating factor

Can G-CSF be implicated in the malignant conversion of congenital neutropenic patients? Development of MDS/AML must be considered in the light of the underlying primary problem. Prior to the availability of G-CSF therapy, it was recognized that leukemic transformation occurs occasionally in patients with congenital neutropenia. In the subgroup of those congenital neutropenic patients with Shwachman syndrome, the predisposition to spontaneous leukemic transformation is inordinately high, possibly up to 30%. However, in the precytokine era, many patients with congenital neutropenia died of other causes in the first years of life. Of published cases, 42% of patients died at a mean age of 2 year secondary to sepsis and pneumonia. Thus, the true risk of patients with congenital neutropenia developing MDS/AML was not defined. Currently with G-CSF therapy, most of these patients do not develop life-threatening infections and are surviving, but it is not known if longer survival will allow for the natural expression of leukemogenesis in this population. Moreover, because the long-term effects of G-CSF are not known beyond 10 years of observation, it is still unclear whether MDS or AML will occur with increased frequency in patients who receive prolonged therapy of G-CSF to correct the neutropenia. So far, this has not been the case, judging from current Registry data quoted herein.

Can G-CSF really be considered carcinogenic? As a physiologic hematopoietic regulator, it would be unexpected for G-CSF to break molecular bonds and cause DNA damage. In well-defined, treatment-related cases of secondary MDS/AML due to chemotherapy for a primary disorder, there is also always a lag phase between the drug exposure and the malignant event. For example, the risk of alkylating agent-related leukemia begins to increase 2 years after start of chemotherapy and peaks in the 5 to 10 year follow-up. Similarly, epipodophyllotoxin-related AML has a median 2 to 3 year induction period following its therapeutic use prior to leukemic conversion. Note in Table 2 that in the 0 to 1 year interval of G-CSF therapy there were three conversions to MDS/AML at varying times within the first 12 months of treatment. Thus, the customary lag phase expected for treatment-induced MDS/AML is not apparent. It seems highly unlikely that G-CSF is directly carcinogenic.

Conclusions

A subgroup of patients with congenital neutropenia who respond well to G-CSF therapy are observed to develop MDS/AML associated with recurrent acquired patterns of cellular genetic changes, singly or in combination (Table 3). Although the exact sequence of events is unclear, the change seemingly begins with a genetic lesion that causes the congenital neutropenia, which then evolves to G-CSF-receptor mutations, monosomy 7, ras oncogene mutations, and finally overt malignant disease.

Carcinogenesis occurs as a series of events that is driven by genetic damage and by epigenetic changes. In the traditional view, the initiation of cancer starts in a normal cell through mutations from exposure to carcinogens. In the promotion phase that follows, the genetically altered, initiated cell undergoes selective clonal expansion that increases the probability of additional genetic damage from endogenous mutations of DNA-damaging agents. Finally, during cancer progression, malignant cells show phenotypic changes, gene amplification, chromosomal alterations, and altered gene expression.

Table 3

Multistep pathogenesis of MDS/AML in severe chronic neutropenia

Genetic Factor
G-CSF therapy
G-CSF receptor mutation
ras oncogene mutation
Monosomy 7
MDS/AML

In congenital neutropenia, the first "hit" or cancer initiating step may be the constitutional genetic abnormality itself that initially manifests as an isolated deficiency of neutrophils. The "predisposed" progenitor, already initiated, could conceptually show decreased responsiveness to the signals that regulate homeostatic growth; for example, less responsive to negative growth factors, terminal cell differentiation, or programmed cell death. The leukemic promotion and progression steps leading to MDS/AML could then ensue readily in the initiated pool of progenitors or stem cells. At this point, we do not know if G-CSF is an innocent bystander in this setting, or whether it participates in the leukemic transformation in some way by accelerating the predisposition in genetically altered congenital neutropenia stem cells.

Currently, G-CSF is still deemed specific therapy for SCN with a high margin of safety and should be the initial treatment for this family of disorders. The issue of malignant conversion with G-CSF therapy is complex in patients with congenital neutropenia who have an underlying propensity for MDS/AML. Thus far, G-CSF has not been directly linked to the transformation events.

Acknowledgements

I am grateful to all my friends and colleagues associated with the Severe Chronic Neutropenia International Registry for valuable information and ideas. Carol Fier and Roger Aitchison were essential for examining the real risk of MDS/AML occurring in SCN patients, and I sincerely thank both of them for their help.