First Line Treatment of Advanced Stage Hodgkin Lymphoma with Six Cycles of BEACOPPescalated Results in Superior Overall Survival Compared to ABVD: Results of a Network Meta-Analysis Including 10,011 Patients

Abstract 551 Background: The best treatment strategy for advanced stage Hodgkin lymphoma (HL) is still a matter of debate. The German Hodgkin Study Group (GHSG) advocates aggressive treatment with BEACOPPescalated (bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine, and prednisone) to cure as many patients as possible with first-line therapy. However, BEACOPPescalated may expose patients to excessive toxicity. Treatment with ABVD (doxorubicin, bleomycin, vinblastine, dacarbazine) is supposed to be better tolerable. Proponents of primary ABVD therapy acknowledge a lower progression-free survival (PFS) compared to BEACOPPescalated. However, they argue that relapsing patients can subsequently be cured by high-dose chemotherapy resulting in comparable overall survival (OS). All trials evaluating these two strategies directly were either very small or included patient subgroups only. Although they congruently showed a significant PFS advantage for BEACOPPescalated, they were not powered to detect differences in OS, which obviously is the most important endpoint. Purpose: To assess the benefits and risks of different initial treatment strategies for adult patients with advanced stage HL and to provide patients and physicians with a high-level evidence for treatment decisions. Methods: Data Sources: We developed sensitive search strategies for CENTRAL, MEDLINE, and conference proceedings (searched from 01/1980 to 03/2012). Missing data was obtained from investigators. Study selection: Randomized trials that compared at least two out of twelve pre-defined chemotherapy regimens in adults with advanced stage HL. Two authors independently assessed studies for eligibility. Data extraction: We extracted data and assessed quality of trials in duplicate. The primary outcome was OS. Secondary outcomes included freedom-from-treatment failure (FFTF) and secondary malignancies. Data relates to four or five years of follow-up depending on the status of the trial. Data synthesis: We pooled data using network meta-analysis. Direct comparisons within trials were combined with indirect evidence from other trials by using a Bayesian random-effects model. Results are reported relative to ABVD with a hazard ratio (HR) >1 indicating superiority of ABVD. Results: 1,984 references were identified, of which 77 publications, reporting 14 trials, evaluating 11 different regimens were included. A total of 10,011 patients with 59,000 patient-years of follow-up were evaluable for the analyses of survival outcomes. Six cycles of BEACOPPescalated and 8 cycles of BEACOPP-14 were associated with the lowest risk for death of any cause (HR 0.38, 95%-CrI 0.20 to 0.75 and HR 0.43, 95%-CrI 0.22 to 0.86, respectively). Assuming a five-year survival rate of 89% for ABVD this would result in a 5-year survival benefit of 7% and 6% for 6 cycles of BEACOPPescalated and 8 cycles of BEACOPP-14, respectively (95%-CrI 3% to 9% and 2% to 9%, respectively). Eight cycles of BEACOPPescalated were also statistically significantly better as compared to ABVD but the effect was less pronounced. All other treatment strategies showed no statistically significant difference to ABVD. Similar results were obtained for FFTF. Between-trial heterogeneity was negligible in both analyses (tau-square 0.01 and 0.05, respectively). Overall, 327 secondary malignancy and 109 leukemia events accumulated over 57,529 patient-years of follow-up. Given the low number of events we were not able to accurately quantify the risk associated with each regimen; however, Stanford V might be associated with the lowest risk and C(M)OPP/EBV/CAD with the highest risk for secondary leukemias. Limitations: Some of the regimens were only evaluated in one trial. The number of secondary malignancies, especially leukemias, was low. Conclusions: The comparison of different first-line treatment strategies for advanced stage HL in this network meta-analysis shows a significant and relevant OS benefit for both, 6 cycles of BEACOPPescalated and 8 cycles of BEACOPP-14 over standard ABVD treatment. This analysis provides the currently best available evidence on OS of different initial treatment strategies for advanced stage HL patients and therefore adds valid and important information for both, patients and physicians. Disclosures: No relevant conflicts of interest to declare.