Effectiveness of rituximab and biosimilar rituximab in untreated diffuse large B-cell lymphoma, a Danish population-based analysis

Diffuse large B-cell lymphoma (DLBCL) is the most common subtype of malignant lymphoma. The annual incidence in Denmark is about 450 cases comparable to other western countries.¹ Conventional chemotherapy, for example CHOP (cyclophosphamide, hydroxydaunorubicin, vincristine and prednisone), has for five decades been the standard treatment of DLBCL.² Rituximab, a monoclonal anti-CD20 antibody drug, was introduced around the year 2000 after positive phase 3 studies in DLBCL.³ Around 2015, the patent of rituximab ended, and biosimilar products were developed. Antibodies are complex molecules consisting of large, folded polypeptide molecules. There is a risk of variation between the molecules, for example due to post-translational modifications. However, the effect of biosimilar rituximab is intended to be identical to generic rituximab. Upon entering the market, the new biosimilar anti-CD20 antibodies (Ritemvia, Rixathon and Ruxience) have been tested in clinical trials against rituximab.^4-7 However, limited data exist on the performance of biosimilar anti-CD20 antibodies in real-world populations.

We aimed to assess the efficacy of biosimilar rituximab compared to rituximab on progression-free survival (PFS), defined as either registered relapse or death from any cause, and overall survival (OS) in Danish patients with previously untreated DLBCL.

Patients with DLBCL diagnosed between January 2015 and January 2022 were identified in the population-based lymphoma registry in Denmark, LYFO.⁸ Inclusion criteria were a first-time diagnosis of DLBCL, treatment with R-CHOP or R-CHOP-like regimens (R-CEOP [cyclophosphamide, etoposide, vincristine and prednisone], R-CHOEP [cyclophosphamide, hydroxydaunorubicin, vincristine, etoposide and prednisone], R-COPE [cyclophosphamide, vincristine, etoposide and prednisone] or R-CHIC [combination of CHOP and CHOEP]). Patients with primary central nervous system lymphoma and missing data on treatment were excluded. The main other regimen than CHOP was R-CHOEP, which comprised 71% in the CHOP-like category. In a recent publication, also based on data from LYFO, outcomes were comparable to standard R-CHOP.⁹

Patients were stratified according to treatment with rituximab or biosimilar rituximab. From the pharmacies of the treating hospitals, we gathered information on the exact date of change to biosimilar rituximab.

To evaluate outcomes, we used the Kaplan–Meier method and Cox regression analysis. The multivariable model was adjusted for International Prognostic Index factors, sex and age. The proportional hazards assumption was evaluated graphically using log minus log plots. Analyses were conducted using Stata/IC 14.2 (Stata Statistical Software: Release 14. StataCorp LP: College Station, TX). The study was compliant with national regulations concerning registry-based research.

We identified 2.372 eligible patients diagnosed with DLBCL between January 2015 and January 2022. They were predominantly males with a median age of 70 years (range: 18–95 years). The median age was 69 years for 1.137 patients treated with rituximab and 71 years for 1.235 patients treated with biosimilar rituximab. The proportion of female patients was 40.7%. No clinically meaningful differences in baseline characteristics were observed (Table 1).

We observed comparable PFS and OS. The 3-year PFS was 71.8% (95% confidence interval [CI] 69.1–74.3) and 72.2% (95% CI 69.3–74.9) for patients treated with rituximab and biosimilar rituximab respectively (Figure 1). The 3-year OS was 77.5% (95% CI 74.9–79.8) and 76.9% (95% CI 74.1–79.4) respectively.

In the multivariable analysis, we found no difference in PFS or OS with a hazard ratio of 0.94 (95% CI 0.81–1.10) and 0.99 (95% CI 0.83–1.17) respectively (Figure 1).

In this study, we evaluated PFS and OS of biosimilar rituximab compared with rituximab in patients with newly diagnosed DLBCL using a national, population-based lymphoma registry in Denmark. The study did not show any difference in OS or PFS. Using nationwide validated data, we were able to identify possible confounders and adjust appropriately. The historical comparison of our two treatment groups occurred within a short time span, and the Danish national treatment guideline for DLBCL was unchanged during that time.

The median age was +2 years in the biosimilar group compared to the rituximab group. This may reflect an increasing willingness over time to offer chemotherapy to elderly patients. Elderly patients typically receive ‘other’ chemotherapy, most likely mini-CHOP. In the rituximab group, treated earlier in the inclusion period, a lower proportion of elderly patients received chemotherapy. Finally, the average age of the general population increased over the study period.

These results are in line with a previous European study.¹⁰ However, in the latter, only OS was reported as an outcome. Improvement of salvage therapies and availability of clinical trials for relapsed/refractory patients in the later years might have improved the OS in the biosimilar group. For this reason, PFS might be a better end-point to assess the effect of biosimilar rituximab.

The main strength of this study is the use of a nationwide population-based lymphoma registry. This allowed for a large sample size. Additionally, the risk of selection bias is minimal due to tax-financed specialist care centralized in few centres. Lastly, uniform nationwide treatment guidelines ensure consistent management across centres. These results support the non-inferiority of biosimilar rituximab, reassuring that use is both safe and efficacious. This could make treatment accessible for more people and result in less economic stress on healthcare systems. These results substantiate the continued use of biosimilar rituximab in patients with DLBCL.

Amal Haujir, Michael Tøstesen, Tarec Christoffer El-Galaly, Thomas Stauffer Larsen, Jørn Starklint, Christian Bjørn Poulsen, Pär Lars Josefsson, Erik Clasen-Linde, Michael Pedersen, Andriette Dessau-Arp, Michael Roost Clausen, and Judit Mészáros Jørgensen all wrote the paper and provided the data possible for this study. Michael Roost Clausen analysed the data. Judit Mészáros Jørgensen designed the study.

No funding was received for this study.

The authors declare no conflicts of interest.

The study was registered at Region Syd, Denmark. Journal nr.: 21/53806.