Michael Fuchs, Horst Müller, Dennis A. Eichenauer, Stefanie Kreissl, Andreas Engert, Peter Borchmann
Department I of Internal Medicine, German Hodgkin Study Group, University Hospital Cologne, Germany
Background: In the treatment of Hodgkin Lymphoma (HL), G-CSF is used to avoid neutropenia and associated complications such as neutropenic fever (FN) and treatment delays when aggressive chemotherapy with escalated BEACOPP (eBEACOPP) is applied. Recently, non-inferiority of a more convenient once-a-cycle application with the pegylated formulation Pegfilgrastim (PF) in comparison with the daily application of Filgrastim (F) and Lenograstim (L) was shown for patients with advanced HL. The present analysis extends this analysis to patients with early unfavorable stages, who were treated with two cycles of eBEACOPP followed by two cycles of ABVD.
Methods: We analyzed the patients treated within the GHSG HD17 trial for the applied G-CSFs, neutropenia, FN and related consequences such as hospitalization. The incidence of FN was the primary endpoint of our study. We applied the Hauck-Anderson method, a type I error probability of 0.05 and a non-inferiority margin of 10% to test the incidence of FN with PF compared to F and L. This primary analysis was constrained to the first two cycles of chemotherapy with eBEACOPP and to patients who had received only one of the three G-CSFs mentioned. We additionally tested whether the application of any G-CSF during the following two cycles of ABVD was associated with a reduced rate of neutropenia.
Results: 1,028 patients received 3,968 chemotherapy cycles; G-CSF was given in 2,969 cycles (1,933 cycles with eBEACOPP and 1,036 cycles with ABVD). During the first two cycles with eBEACOPP mostly PF was used (69.6% = 1,346/1,933) followed by F (12.1% = 233/1,933), L (5.6% = 109/1,933) and other G-CSFs (10.3% = 200/1,933). Altogether, the incidence of FN was 14.3% during the cycles with eBEACOPP; 10.4% of patients were hospitalized due to FN. With PF the incidence of FN was 13.5% (=94/699), with F 17.0% (=19/112), and with L 20.8% (11/53). Non-inferiority of PF to prevent FN was demonstrated when compared to F (P = .0003) and L (P = .0023). The percentage of hospitalizations because of FN was 10.0% (=70/699) with PF, 15.2% (=17/112) with F and 18.9% (=10/53) with L. During cycles three and four with ABVD, the incidence of FN using any G-CSF was 3.0% (=23/768) vs. 5.5% (=11/200) without use of G-CSF (P = 0.09).
Conclusion: In patients with early unfavorable HL, PF is at least as effective as F and L to prevent FN during the first two cycles with eBEACOPP. The use of G-CSFs during the following two cycles with ABVD may prevent FN.