Hodgkin lymphoma – absence of evidence not evidence of absence!

Peter Borchmann, Andreas Engert and Volker Diehl

The optimal treatment for patients with advanced-stage Hodgkin lymphoma is an ongoing controversy. A recent trial seemed to answer some of the important open questions in the field; however, closer examination of the data indicates that the answers are not as clear as they might initially seem.

The majority of patients (85–95%) with advanced-stage Hodgkin lymphoma can be cured; however, it is currently unclear which treatment strategy offers the best balance between toxicity and efficacy. This key question has been discussed extensively since the introduction of the dose-intense therapy BEACOPP (bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine and prednisone) a decade ago. The efficacy and toxicity of BEACOPP in patients with advanced-stage Hodgkin lymphoma was initially evaluated in the GHSG-HD9 trial that compared it with COPP–ABVD (an alternating regimen of cyclophosphamide, vincristine, procarbazine and prednisone [COPP], and doxorubicin, bleomycin, vincristine and dacarbazine [ABVD]).[1] With 10 years of follow up, the escalated regimen of BEACOPP was clearly superior to COPP–ABVD in terms of tumour control – the difference in freedom from treatment failure was 18% and the difference in overall survival was 11%.[2] However, despite the efficacy superiority of BEACOPP over COPP–ABVD, ABVD alone followed by consolidation radiotherapy for residual disease (required by approximately 65% of patients) is still commonly being used to treat patients with advanced-stage Hodgkin lymphoma. This lack of change in the treatment paradigm has arisen because COPP–ABVD was not regarded as standard of care by some opinion leaders owing to its similar efficacy but increased toxicity compared with ABVD alone.[3,4]

With regard to the lower efficacy of ABVD as compared with BEACOPP (five-year progression-free survival difference 14%; overall survival difference 8%),[5] those advocating ABVD as first-line treatment refer to the ‘second-shot’ hypothesis that includes high-dose chemotherapy as late intensification in patients who have progressive disease or who relapse after first-line ABVD. These patients can be rescued with high-dose chemotherapy – the ‘second shot’. This strategy results in an overall survival rate of 80–85% for the whole group, including those patients (estimated 25–35%) who require high-dose chemotherapy rescue.

By contrast, an early intensification approach (the ‘Kairos Principle’) using the more-effective (but also more-toxic) BEACOPP regimen aims to cure as many patients as possible with a ‘first shot’. After treatment with BEACOPP, only 12% of patients need consolidation radiotherapy and only 10–15% of patients will relapse,[6] which results in an overall survival rate of 90–95% at five years (A Engert et al., unpublished data). First-line treatment with BEACOPP has been adopted by most European study groups as standard-of-care for advanced-stage Hodgkin lymphoma.

When one considers particularly the survival difference between treatment with ABVD and BEACOPP one might wonder why one should treat young and otherwise fit patients with Hodgkin lymphoma using ABVD. The reason for the persistent use of ABVD in this patient population is the lack of evidence from randomised trials. Only one trial has directly compared BEACOPP with ABVD, and this trial was too small to obtain significance for survival rates.[5] Therefore, there has been a debate on the question of whether ABVD results in preventable death for one in ten patients or if BEACOPP is an unnecessarily aggressive overtreatment for three quarters of patients.[6]

Against the background of this controversial discussion, a recent publication in the New England Journal of Medicine has come as a surprise. Viviani et al.[7] report a direct comparison of a modified BEACOPP regimen with ABVD; both regimens were followed by salvage high-dose chemotherapy for relapsing or progressing patients. The aim of this study was to analyse long-term disease control and treatment-induced morbidity.

Although the comparison of these treatment strategies had been eagerly awaited, unfortunately the trial design is inadequate to answer this ‘main’ objective. The trial was designed and powered only for the primary endpoint (freedom from first progression), and testing hypotheses for secondary endpoints such as overall survival is specifically excluded by the authors (see statistical analysis plan, page 14, section 14.3 in the online appendix).7 Nonetheless, P values for secondary endpoints are presented throughout the manuscript. The 5% overall survival difference in favour of BEACOPP is interpreted by the authors as ‘nonsignificant’ (P=0.39) and an overall benefit for the less-toxic ABVD treatment is concluded; actually, the lack of significance is a limitation of the study, not a result. In fact, the results of this trial are in sharp contrast to the authors’ conclusion because there was a survival difference in favour of BEACOPP in line with the significantly superior freedom from first progression, which was the primary endpoint (P=0.004).[7] In addition, the conclusions are in contrast to published data showing an overall survival benefit of 8–11% associated with treatment with BEACOPP.2,5 Thus, if one considers overall survival to be the most relevant endpoint for young patients with cancer, the data presented by Viviani et al.[7] rather support the early intensification approach provided by first-line treatment with BEACOPP.

When making the assessment of which therapy to use in the first-line setting, one should carefully weigh potential risks and benefits and should have a closer look at the data reported. This closer look at the data is, unfortunately, disappointing; not only the design but also the reporting quality of the Viviani et al.7 study is surprisingly deficient. The manuscript contains numerous mistakes and discrepancies, both within the publication itself and when compared with a previously reported interim analysis of the same trial, as we have outlined in a letter to the authors that has been accepted for publication in the New England Journal of Medicine.[8] Furthermore, regarding the important comparison of the toxic effects associated with the two treatments, Viviani et al.[7] emphasise the problem of late toxicities induced by BEACOPP, but they observed three secondary malignancies in the BEACOPP group, and four in the ABVD group. To underline the good tolerability of ABVD, they cite the results from a clinical trial in which only six cycles of ABVD were applied to each patient;[9] however, eight ABVD cycles were administered in their own study. In fact, little is known about the long-term sequelae (for example, infertility, therapy-induced cardiac dysfunction, bleomycin-induced pulmonary toxicity, and quality of life) associated with eight cycles of ABVD. Unfortunately, the authors have missed the opportunity to add important and missing information to the field.

Even though the publication of the data from the trial seems to have some serious limitations, the reported statistic that 73% of patients might be cured using ABVD as first-line treatment should be accepted and interpreted. The important question then is how to safely discriminate at diagnosis those patients who will respond to ABVD from those who will progress or relapse after ABVD therapy, which unfortunately is still impossible. As long as we cannot detect these patients, who have a difficult-to-treat and life-threatening disease, patients and physicians should be aware of the overall survival difference in favour of BEACOPP for patients with advanced-stage Hodgkin lymphoma. The conclusion that a 5–11% survival difference is not relevant might sound strange or even cynical to some patients.

Needless to say, we should be carefully looking for evidence to provide the best treatment for our patients. Therefore, we rely on properly designed, conducted and reported clinical studies. Fortunately, medicine is more than just politics. With this in mind, we look forward to the upcoming publication of the EORTC 20012 trial comparing ABVD with a modified regimen of BEACOPP in patients with advanced-stage Hodgkin lymphoma.10 Even more so, we need to enrol patients into worldwide ongoing clinical trials investigating new strategies such as PET-guided response-adapted treatment to restrict more-aggressive treatment to the subset of patients who really need it.


1. V Diehl et al. (2003) Standard and increased-dose BEACOPP chemotherapy compared with COPP-ABVD for advanced Hodgkin’s disease. NEJM 348:2386–95

2. A Engert et al. (2009) Escalated-dose BEACOPP in the treatment of patients with advanced-stage Hodgkin’s lymphoma: 10 years of follow-up of the GHSG HD9 study. JCO 27:4548–54

3. GP Canellos et al. (1992) Chemotherapy of advanced Hodgkin’s disease with MOPP, ABVD, or MOPP alternating with ABVD. NEJM 327:1478–84

4. DB Duggan et al. (2003) Randomized comparison of ABVD and MOPP/ABV hybrid for the treatment of advanced Hodgkin’s disease: report of an intergroup trial. JCO 21:607–614

5. M Federico et al. (2009) ABVD compared with BEACOPP compared with CEC for the initial treatment of patients with advanced Hodgkin’s lymphoma: results from the HD2000 Gruppo Italiano per lo Studio dei Linfomi Trial. JCO 27:805–811

6. P Borchmann and A Engert. (2010) The past: what we have learned in the last decade. Hematology Am Soc Hematol Educ Program pp 101–107

7. S Viviani et al. (2011) ABVD versus BEACOPP for Hodgkin’s lymphoma when high-dose salvage is planned. NEJM 365:203–212

8. V Diehl et al. (in press) Should the treatment of advanced Hodgkin lymphoma patients be a question of faith or a question of medical science? NEJM

9. A Santoro et al. (1987) Long-term results of combined chemotherapy-radiotherapy approach in Hodgkin’s disease: superiority of ABVD plus radiotherapy versus MOPP plus radiotherapy. JCO. 5:27–37

10. US National Library of Medicine. (2010) ClinicalTrials.gov

Author affiliations: German Hodgkin Study Group, University Hospital of Cologne, Cologne, Germany (Peter Borchmann, Andreas Engert and Volker Diehl)

This article was first published in Nature Reviews Clinical Oncology vol.8 no.11, and is published with permission. © 2011 Nature Publishing Group. doi:10.1038/nrclinonc.2011.149,


Be the first to comment

Leave a Reply

Your email address will not be published.