Effects of antibiotic prophylaxis in first-line therapy of advanced stage classic Hodgkin lymphoma: An analysis of the GHSG HD21 study

Abstract Introduction: The benefit of antibiotic prophylaxis (ABP) during chemotherapy for cancer remains controversial. The growing resistance to fluoroquinolones in hematological patients (Gupta et al., Lancet Oncology, 2025) and concerns about adverse events challenge its widespread use. In the GHSG HD21 trial for advanced-stage classic Hodgkin lymphoma (AS-cHL), patients received polychemotherapy regimens (eBEACOPP [bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine, and prednisone] or BrECADD [brentuximab vedotin, etoposide, cyclophosphamide, doxorubicin, dacarbazine, and dexamethasone]) with a risk to develop febrile neutropenia (FN) and/or infections. Here, we investigate the impact of ABP and Granulocyte colony-stimulating factor (G-CSF) formulation (pegylated vs. non-pegylated) on FN and infection rates. Methods HD21 is an international, prospective, open-label, randomized trial of individualized PET2-guided treatment of AS-cHL with either eBEACOPP or BrECADD. In this post-hoc analysis patients, who received at least one cycle of treatment, were included. ABP was recommended per protocol but not mandatory. G-CSF support was mandated. Associations between ABP use and patient, center and treatment characteristics were tested via χ²/Fisher's exact and unequal-variance t-tests. Rates of FN and CTCAE Grade (G) ≥3 infections were compared per cycle. The effects of ABP and pegylated (peg) vs non-peg G-CSF were analyzed separately over all cycles available. Due to high overall survival rates for both regimens, we were not able to assess the effects of ABP on mortality. Results 1.500 patients were enrolled into the HD21 trial, of whom 1470 patients (with N=6955 cycles) received at least one cycle of treatment (ITT-TRMB). Of these patients, 984 (67 %) received ABP in at least one cycle. Per cycle, usage of fluoroquinolone ABP ranged between 97.1% and 98.1%. ABP was more commonly used in female than male patients (71.1 % vs. 63.6 %, p=0.0026) and at academic sites and outpatient clinics than in non-academic hospitals (71.0% vs. 82.9 % vs. 61.9 %, p<0.0001). The proportion of patients experiencing FN or infection ≥ G3 was higher in the first (15.9%; 9.2%) than in subsequent cycles (4.7-7.6%; 4.1-5.7%) and in this cycle especially high for patients ≥40 years (18.4%; 12.3%) vs <40 years (15.0%; 8.0%). We found a benefit for the use of ABP with BrECADD: The rate of FN was 9.8 % without ABP and 7.5 % with ABP (p = 0.019), and for ≥ G3 infections 6.7 % and 4.5 % (p = 0.0065), respectively. With eBEACOPP, there were no significant differences (FN: 6.8 % vs 6.3 %, p = 0.54; ≥ G3 infections: 5.6 % vs 4.9 %, p = 0.40). Focusing on the first cycle, reduction was present with ABP in both arms: With eBEACOPP (n=732), FN occurred in 16.1 % vs 9.7 % and ≥ G3 infections in 9.7 % vs 6.5 %; with BrECADD (n=738), occurrence of FN was 22.0 % vs 16.6 % and ≥ G3 infections 14.0% vs 7.1 %. Non-peg G-CSF was used in 599 (8.7 %) of 6905 cycles. With BrECADD, the incidence of FN (8.3% vs. 13.8%, p = 0.0051) was significantly lower with peg G-CSF vs non-peg G-CSF, whereas no such difference was observed with eBEACOPP (6.4% vs. 8.4%, p = 0.17). Conclusions This comprehensive analysis of the HD21 trial demonstrates the efficacy of ABP and peg G-CSF in the treatment of AS-cHL in preventing FN and higher-grade infections. This effect is most pronounced in the first cycle and in patients receiving BrECADD. Based on these results, we recommend the use of ABP during the first cycle of the BrECADD regimen, at least.