Granulocyte colony-stimulating factor (G-CSF)

Observational studies

In all clinical studies carried out to date, G-CSF has been well tolerated, whether given subcutaneously or intravenously. At the recommended doses (5–10 micrograms/kg), generalized musculoskeletal and transient bone pains, headache, and mild rash are the commonest adverse effects [ ]. No additional adverse effects or delayed consequences have been so far reported in neonates treated at birth for presumed bacterial sepsis [ ]. An increase in the size of the spleen has been reported [ , ]. Transient rises in alkaline phosphatase, lactate dehydrogenase, and uric acid are considered to be normal physiological consequences of the rise in the neutrophil count [ ]. Long-term G-CSF administration in patients with severe congenital neutropenia has also been considered to be relatively safe, with discontinuation or temporary withdrawal in only seven of 44 patients [ ].

Two reviews have examined the available data on the effects of hemopoietic growth factors on the duration of neutropenia and mortality in drug-induced agranulocytosis, which mostly consists of isolated case reports or small series of patients [ , ]. The authors reached contrasting opinions, suggesting that hemopoietic growth factors might or might not be of interest in patients with severe drug-induced agranulocytosis. Adverse effects were noted in 13 of 118 case reports [ ]. Although most of them were benign, pulmonary toxicity or acute respiratory distress syndrome have been noted in a few patients.

The Severe Chronic Neutropenia International Registry (SCNIR) has collected data from patients with hematological disorders causing blood neutrophil counts below 500 × 106/l for months or years [ ]. Of 1163 patients in the SCNIR 1053 (91%) have been treated with granulocyte colony-stimulating factor (G-CSF, median dose 3.33 μg/kg/day). Adverse effects included osteoporosis/osteopenia (14%) and myelodysplastic syndrome or acute myelocytic leukemia (4.9%).

Comparative studies

Few studies have directly compared the safety of the various available colony-stimulating factors. The frequency and severity of adverse effects associated with the prophylactic use of filgrastim (a bacterial cell-derived G-CSF) or sargramostim (a yeast cell-derived GM-CSF) have been assessed in a retrospective review of the medical records of 490 cancer patients from 10 centers [ ]. Sargramostim-treated patients had significantly more frequent non-infectious fever, fatigue, diarrhea, injection site reactions, edema, and dermatological adverse effects, whereas skeletal pain was more frequent with filgrastim. In addition, switching to the alternative treatment was more frequent in the sargramostim group (18% of patients) than in the filgrastim group (none of the patients). The authors tried to minimize selection bias, but the strength of the results was limited by the retrospective nature of the study.

The effects and the safety of a 5-day regimen of G-CSF (n = 9) or GM-CSF (n = 8) have been compared [ ]. Most patients complained of flu-like symptoms in both groups (six and seven respectively), but rash at the injection site was observed only in four patients treated with GM-CSF. In the G-CSF group, there was a fall in platelet count (below 150 × 10 ⁹ /l) in five patients, raised serum lactic dehydrogenase activity, and raised uric acid concentrations; three patients required transient treatment with allopurinol.

In a preliminary randomized, blind, controlled comparison with usual care in 10 patients with acute cerebral infarction who presented within 7 days of onset, subcutaneous G-CSF 15 micrograms/kg/day for 5 days produced greater improvement in neurological function; there were no severe adverse effects [ ]. However, concerns about the use of G-CSF in these circumstances have been raised because of fears about transient hypercoagulability, which may be important for thrombophilic, atherosclerotic, or immobilized patients [ ]. This may be related to increased amounts of endothelial markers and thrombin generation or by stimulating tissue factor. There have been a few reports of acute arterial thrombosis in patients receiving G-CSF, supporting the hypothesis of hypercoagulability. In addition, there have been cases of acute arterial thrombosis in healthy donors, possibly related to G-CSF.

Use in healthy volunteers

The G-CSF is sometimes used in healthy volunteers to mobilize blood progenitor cells or granulocytes before infusion into neutropenic patients [ ].

The safety of filgrastim in healthy donors has been evaluated in a large prospective multicenter study [ ]. The interim results, obtained from the first 150 enrolled donors aged 18–64 years who received either 10 or 16 micrograms/kg/day, have shown that 99 patients had at least one adverse effect graded as mild (grade I) in 35% of cases, moderate (grade II) in 62%, and severe (grade III) in 3%. Bone pain and headaches were the most common acute adverse events, and all the patients completely recovered after withdrawal of filgrastim. There were no apparent differences in the proportion, the severity, or the type of adverse effects according to the administered regimen.

That the use of G-CSF in healthy donors of peripheral blood progenitor cells is reasonably safe has been confirmed in an analysis of adverse effects in 737 evaluable patients included in three independent databases from Spain, the USA, and Japan [ ]. In one study, the overall incidence of adverse effects was 67%. The most common adverse effects were bone pain (71–90%), headache (17–54%), fatigue (6–33%), insomnia (up to 14%), nausea/vomiting (3–13%), and low-grade fever (6%). Although most adverse effects were rated as moderate, about two-thirds of the patients required analgesics for bone pain or headache. Other adverse events, such as non-cardiac chest pain, paresthesia, itching, or minor injection site reactions, were rare. Very few patients discontinued G-CSF because of clinical toxicity. Doses higher than 8.8 micrograms/kg/day, patients younger than 35 years of age, and female sex were significant risk factors for bone pain, headache, and nausea/vomiting respectively [ ].

Moderate thrombocytopenia was common after apheresis; there was more severe but asymptomatic and promptly reversible thrombocytopenia (below 50 × 10 ¹² /l) in up to 3.9% of patients [ , ]. However, the respective contributions of apheresis and G-CSF in thrombocytopenia are difficult to assess. There was also a fall in the absolute neutrophil count to less than 1 × 10 ⁹ /l in 3% of patients 9–16 days after G-CSF withdrawal [ ]. Serum alkaline phosphatase and lactate dehydrogenase activities were increased about two-fold compared with pre-treatment, a finding that was explicable by G-CSF-induced neutrophilia [ , ]. Symptoms of hypercoagulability have also been found in healthy donors.

However, more severe or unexpected consequences, including spontaneous splenic rupture, anaphylactoid reactions, deep necrotizing folliculitis, a psoriasiform eruption, acute gouty arthritis, iritis or episcleritis, and unexpectedly prolonged thrombocytopenia have been mentioned [ ].

Among 90 healthy donors given 10 and 16 micrograms/kg/day of filgrastim for stem cell mobilization, severe adverse effects were mostly found in patients who had been given the higher dose [ ], but one obese patient (body weight 170 kg) who received 10 micrograms/kg/day had a non-traumatic spleen rupture that resolved spontaneously.

Sufficient data on the potential long-term effects of G-CSF in healthy volunteers are still lacking. In one study, 101 healthy donors who had received filgrastim for a median of 6 days were questioned after a median of 43 (range 34–74) months to assess their current health; 70 donors also had a complete blood count [ ]. No unusual disease was detected and the blood counts were within the reference range. In 20 donors followed for 6–12 months after peripheral blood progenitor cell collection, there were no particular symptoms or hematological abnormalities [ ].

The possible carcinogenic effects of G-CSF have also been discussed. There is as yet no indication of an increased risk of cancer, and no cases of acute or chronic leukemia were detected in a telephone interview study performed after a median of 39 months after peripheral blood progenitor cell donation in 281 donors [ ].

Cardiovascular

Cardiovascular events have seldom been described in patients given colony-stimulating factors. However, possible excesses of cardiovascular events and unexpected deaths have been suggested [ ], although the actual risk was not fully evaluated. In three isolated reports, acute arterial thrombosis or angina pectoris were deemed to have resulted from hypercoagulability with extreme leukocytosis and G-CSF-induced abnormalities in platelet aggregation [ ]. Increased platelet aggregation has also been found in healthy volunteers [ ]. Although the relevance of these findings is unclear, caution is warranted in patients predisposed to thromboembolic events.

• A 46-year-old donor denied pre-existing cardiac symptoms, but smoking and a family history of coronary artery disease were noted as possible risk factors [ ]. The pre-treatment electrocardiogram was normal. Six hours after the second and the third doses of G-CSF 10 micrograms/kg before peripheral blood progenitor cell collection, he developed symptoms and signs of cardiac ischemia, including palpitation, chest discomfort, trigeminy, and T wave inversion. However, troponin was unchanged. Cardiac catheterization showed severe coronary artery occlusion and he underwent percutaneous transluminal coronary angioplasty. He finally admitted mild exertional chest discomfort 2 weeks before the first dose of G-CSF.

Although not described during clinical trials, typical capillary leak syndrome has been anecdotally observed after G-CSF administration, illustrating the possible consequences of accelerated release of activated granulocytes [ ].

Microthrombotic necrotizing panniculitis has been reported [ ].

• A 49-year-old woman received subcutaneous filgrastim 300 micrograms/day into the upper thighs for neutropenia prophylaxis after treatment of relapsing Hodgkin’s disease with mitoguazone, etoposide, vinorelbine, and ifosfamide. After 3 days she suddenly developed fever, painful livedo, deeply infiltrated edema on the legs and thighs, and inflamed livedoid erythema on both soles. Deep biopsy specimens showed small vessel thrombosis with subcutaneous necrosis and hemorrhage. She recovered over the next 4 weeks after filgrastim withdrawal and prednisone treatment.

Although a causal relation was difficult to ascertain in the context of malignancy and cytotoxic chemotherapy, the short time to occurrence after G-CSF favored a causative role.

Respiratory

Nervous system

There is no clear evidence of specific central nervous system adverse effects due to G-CSF. In one patient, neurological symptoms, such as blurred vision, weakness, and headache, were attributed to G-CSF-induced an extreme leukocytosis with subsequent hyperviscosity [ ]. Encephalopathy, cortical blindness and seizures have also been mentioned in a single case report [ ].

Sensory systems

Severe retinal hemorrhage with a slowly reversible loss of visual acuity, and massive vitreous hemorrhage recurring after further G-CSF treatment and resulting in irreversible loss of vision in the affected eye have each been reported in single patients [ , ]. Concomitant leukocytosis was suggested as a possible cause in the first patient, whereas G-CSF-induced reactivation of primary ocular inflammation (probably infectious in origin) was advanced as an explanation in the second case.

• A 61-year-old healthy donor developed marginal keratitis with associated mild uveitis after being given injections of filgrastim and sargramostim for 3 days [ ]. Topical prednisolone and withdrawal of sargramostim produced improvement within 24 hours, while filgrastim injections were continued.

Iritis and episcleritis have also been mentioned in healthy donors of blood progenitor cells.

Endocrine

Several studies have suggested that single or short-term administration of G-CSF did not produce significant changes in the serum concentrations of cortisol, growth hormone, prolactin, follicle-stimulating hormone, luteinizing hormone, or thyrotropin [ , ].

Thyroid function and thyroid antibodies were not modified in 20 breast cancer patients [ ], and only one case of hypothyroidism with increased thyroid antibodies has been reported [ ]. G-CSF had no effect on thyroid function in 33 patients with cancer, even in patients with pre-existing antibodies [ ]. Subclinical and spontaneously reversible hyperthyroidism occurred in eight patients without thyroid antibodies and with normal thyroid function before treatment, but this was felt to be related to stressful procedures.

Metabolism

Reductions in serum cholesterol concentrations have been sometimes noted in patients receiving G-CSF [ ].

Hematologic

The long-term effects of in vivo G-CSF stimulation of the hemopoietic system during radiotherapy have been investigated in a randomized trial in 10 patients (mean age 51 years; 6 men, 4 women) with non-Hodgkin’s lymphoma (n = 7), Hodgkin's lymphoma (n = 1), or small-cell carcinomas of the lung (n = 2) [ ]. They were randomized to either large-field radiotherapy alone or radiotherapy with simultaneous G-CSF and were assessed for acute and late toxicity. Platelets and CD34 superset + progenitor cells fell significantly in the G-CSF treatment group and further treatment was stopped. Peripheral leukocyte counts were 2.8–4.4 × 10 ⁹ /l in 9/10 patients. In those who were given G-CSF mean platelet counts fell below 30 × 10 ⁹ /l and CD34 superset + progenitor cells to 50%. Hemoglobin did not vary. Differential blood smears showed differences in granulocyte counts and a higher proportion of neutrophils in those who were given G-CSF. Lymphocyte counts in patients who were not given G-CSF were significantly less than in those who were. Simultaneous treatment with G-CSF during radiotherapy reduced the mobilization of CD34 + progenitor cells and exhausted the bone marrow capacity, while peripheral leukocyte counts remained at baseline.