Patterns of Mortality After Prolonged Follow-up of a Randomised Controlled Trial Using Granulocyte Colony-stimulating Factor to Maintain Chemotherapy Dose Intensity in Non-Hodgkin’s Lymphoma

Br J Cancer. 2008 Aug 22;99(2):253-258, AR Clamp, WD Ryder, S Bhattacharya, R Pettengell, JA Radford

Myelosuppression and increased susceptibility to infection limit the dosages of cytotoxic chemotherapies used in patients with non-Hodgkin’s lymphoma (NHL). To reduce the incidence of these toxicities and thereby maintain dose intensity of chemotherapy, patients often receive adjunctive therapy with hematologic colony-stimulating factors, such as granulocyte colony-stimulating factor (G-CSF). A meta-analysis of previous studies of G-CSF use in patients with NHL showed a 26% reduced risk of febrile neutropenia. However, no overall survival benefit was observed. Because the median follow-up in the meta-analysis was only 4.4 years, UK investigators analyzed long-term follow-up data from the initial randomized trial of G-CSF in NHL patients to determine how the use of G-CSF and the resulting increase in chemotherapy intensity affects long-term survival, overall morbidity, and late complications. They found that treatment with G-CSF was associated with a greater risk of death due to non–NHL-related causes, including infections, second malignancies, and cardiovascular disease. These deaths may be related to increased exposure to chemotherapy permitted by the use of G-CSF or to G-CSF itself. Thus, the relationship of G-CSF to these causes of death should be investigated further.

Of 80 patients with high-grade NHL treated with VAPEC-B (vincristine, doxorubicin, prednisolone, etoposide, cyclophosphamide, and bleomycin) chemotherapy, 41 were randomized to receive G-CSF 230 mcg/m² administered subcutaneously daily for 13 weeks and 39 received chemotherapy alone. Median dose intensity of chemotherapy was increased by 12% for patients receiving G-CSF compared with control subjects.

Of the 80 original patients, 48 died. Median follow-up was 15.7 years in the surviving patients (range, 8.4-16.9 years). The 10-year overall survival (OS) rates were 51% and 46% in the G-CSF and control groups, respectively; the 10-year progression-free survival (PFS) rates were 49% and 44%, respectively. The 10-year freedom from progression (FFP; time from randomization to first documentation of disease progression) rates were significantly higher in the G-CSF group than in the control group (68% vs 47%, respectively). In addition, 11 (27%) and 21 (54%) patients in the G-CSF and control groups, respectively, had progressive disease, of whom 10 and 21 patients died, respectively (P = .02). Notably, 14 and 5 patients in the G-CSF and control groups, respectively, died of causes other than NHL. Patients receiving G-CSF were more likely to die before lymphoma progression (hazard ratio [HR] = 3.08; 95% confidence interval [CI], 1.05-8.99) and were less likely to experience lymphoma progression (HR = 0.40; 95% CI, 0.18-0.87) than control subjects. In the G-CSF group, non–NHL-related deaths included treatment-related infections (n = 3), neutropenic infections (n = 2), and invasive aspergillosis (n = 1). In contrast, no control subjects died as the result of treatment-related infections. Twice as many patients in the G-CSF group than in the control group died of second malignancies (4 vs 2 patients). The remaining 6 deaths in the G-CSF group included 5 due to cardiovascular disease and 1 due to intracerebral hemorrhage. The remaining 2 deaths in the control group included 1 death from Pneumocystis  carinii  pneumonia, which occurred 14 months after the patient received chemotherapy and was not associated with recurrent NHL, and 1 death from cirrhosis secondary to chronic hepatitis C infection. Relative mortality models based on general population death rate data from England and Wales suggested that, after the period of early excess relative deaths associated with lymphoma, the death rate among surviving patients was higher in the G-CSF group than in the control group. 

These data suggest that G-CSF may improve lymphoma-specific survival outcomes but may also increase the likelihood of death from other causes, including infections, second malignancy, and cardiovascular disease. Further investigation is needed to determine the extent to which the increase in deaths due to these non-NHL causes may be related to increased dose intensity of chemotherapy or to G-CSF itself.

View this article via PubMed »

Summary written by the OncologySTAT editorial team