Study of Neupogen vs Biosimilars Draws Fire

After a study from India suggested a higher incidence of adverse reactions in patients using biosimilar filgrastims, 2 European biosimilar consultants and oncologists raised objections to the study’s methodology.

When Shruti Rastogi, MPharma, and her fellow investigators published a study suggesting that biosimilar filgrastims have a higher incidence of certain adverse events than the originator product, Neupogen, they immediately drew fire from 2 established experts on biosimilar development, Paul Cornes, MD, and Michael Muenzberg, MD, both of the Comparative Outcomes Group.

Cornes and Muenzberg found what they described as deep flaws in the methodology used by the investigators as well as the real-world database they relied on, VigiBase, which Cornes and Muenzberg said could not reliably be used to generate the type of conclusions reached by Rastogi and her fellow investigators.

The core problem, they said, was that the study unjustly portrays multiple filgrastim biosimilars as inferior to the originator product without a truly scientific examination of the data. They have asked the journal involved, Toxicology and Applied Pharmacology, to at least print a balancing statement that would expose some of the concerns about the study’s validity.

In response to a query from The Center for Biosimilars®, Rastogi defended the findings as valid with a point-by-point response to the concerns identified by Cornes and Muenzberg and contended that the study was unbiased and should serve as a starting point for further studies to investigate the findings. The “authors have emphasized that this stimulating study is just a guide for further evaluation regarding safety and comparability of filgrastim originator and biosimilars,” Rastogi said. “We have no intention to discredit biosimilars use and bolster consumer loyalty to originator products.”

“Higher” Incidence of Adverse Reactions

Specifically, the study indicates that compared with Neupogen, the biosimilar filgrastim Grasin was associated with higher reporting of pyrexia (11.5% vs 7.9%), myalgia (37% vs 2.2%), and back pain (11.3% vs 4%). Zarzio was associated with higher reporting of arthralgia (4.5% vs 2.9%) and neutropenia (11.4% vs 4%), and Nivestim was associated with a higher incidence of bone pain (14.4% vs 8.3%). The study reports instances of ineffectiveness for the biosimilars Zarzio, Nivestim, and Tevagrastim, but not for Neupogen. Further details of the study are available here.

Cornes and Muenzberg, both oncologists and investigators, have credentials. They have consulted for all of the companies involved with the originator product and biosimilars included in the study. Muenzberg is former medical director of EU biosimilars for Amgen, the company that developed and markets the reference product Neupogen. Cornes presented evidence to the FDA for the first approved US biosimilar.

One of their chief criticisms of the study is that it relies on comparisons made using the VigiBase adverse event database from the World Health Organization (WHO), which WHO specifies is appropriate for “generating but not for testing hypotheses” because the data it contains are subject to “reporting biases, duplication, confounding issues,” in which variables distort outcomes, and other problems.

Rastogi, lead author of the study and an assistant scientist at the Indian Pharmacopoeia Commission, contended the large pool of global information in the WHO database facilitated the detection of drug safety signals and increased “the probability of detecting rate adverse drug reactions.” She said similar studies have been done using global pharmacovigilance databases like VigiBase.

“Incorrect” Use of P Numbers

Cornes and Muenzberg also found fault with the study’s application of P numbers, which express statistical significance, to the percentage differences given for adverse events connected with biosimilars in the study. They said use of equivalence margins, which specify the range of acceptability or nonacceptability for biosimilarity to the originator product, are standard procedure for such comparative studies and would have provided meaningful results.

Rastogi responded that use of P numbers helped to determine potential adverse drug reaction associations between originator and biosimilars based on pharmacological hypotheses. “The authors have not drawn any conclusion based on P values,” she said. “In our study we were able to demonstrate that Zarzio exposure was significantly associated with a high reporting odds ratio (ROR) value (1.59) for arthralgia as compared to innovator with ROR (1.21) using VigiBase.”

Cornes and Muenzberg also contend that the investigators wrongly declared that there is a serious lack of real-world, postmarketing data on biosimilar safety. Cornes and Muenzberg said that 1500 patients have been enrolled in head-to-head randomized trials of the filgrastim biosimilars mentioned by Rastogi and her fellow investigators. Thousands more patients have been enrolled in postmarket safety studies of filgrastim biosimilars, they said.

Rastogi responded that biosimilar developers’ lack of access to manufacturing process data for the originator product creates a “high possibility” for minor structural differences in biosimilars that may “result in a product that differs in terms of pharmacological specifications, efficacy, and safety including immunogenicity influencing the product’s benefit-risk profile.” These differences may be missed by studies that demonstrate biosimilarity to the originator product, Rastogi said.

Rastogi further defended the integrity of the study, contending that it was subject to an extensive peer review by the journal Toxicology and Applied Pharmacology. Cornes and Muenzberg contended the study would have flunked what they described as an appropriately rigorous peer review.

A final criticism from Cornes and Muenzberg was that in their study the investigators appeared to inappropriately lump together biosimilars with India-approved copy biologics, which are certified for use by a different set of standards. The authors’ analysis and conclusions appear to be derived from this combination of dissimilar drugs, Cornes and Muenzberg said. However, although the study makes mention of copy biologics, those were not a part of the final analysis and conclusions, Rastogi said.

Cornes and Muenzberg said that they are not “unsympathetic” to the need to generate additional data on the safety and efficacy of biosimilars, but they contend that the findings outlined in the study are “unsustainable” and contrast with existing data provided in Periodic Safety Update Reports, which are risk-benefit pharmacovigilance reports provided by biosimilar developers at intervals following marketing authorization, and these are relied upon by regulators in their continuing assessments of biosimilars.

The controversial aspects of the study have attracted notice. Julie Maréchal, director of Biosimilar Policy and Science for Medicines for Europe, an equal access advocacy group, said that the accumulated evidence on biosimilars in the European Union has sufficiently demonstrated that these are safe medicines. “The European Biosimilar Medicines Group, a Medicines for Europe sector group, concurs with expert opinion expressed on the merits of the Rastogi et al paper. The facts and their interpretation do not fully match nor take account of important limitations. There is no question: EU-authorised biosimilar medicines have a proven safety track record,” Maréchal said.

In response to an inquiry from The Center for Biosimilars®, Pfizer commented that the study findings do not weaken its faith in its own filgrastim biosimilar, Nivestym. “We are confident in the safety and efficacy profile of our filgrastim biosimilar, and the comprehensive data package and totality of evidence which demonstrates a high degree of similarity of this medicine compared to its reference product.”

Amgen, which distributes Neupogen and could potentially benefit from the study owing to the conclusion that Neupogen is superior to its biosimilars based on adverse drug reactions, did not respond to a request for comment.

In the study, Rastogi and fellow investigators state that they had no conflicts of interest to report and that their funding source was not involved in study design or data collation, analysis, or interpretation.

Reference

Rastogi S, Shukla S, Sharma AK, et al. Towards a comprehensive safety understanding of granulocyte-colony stimulating factor biosimilars in treating chemotherapy associated febrile neutropenia: trends from decades of data. Toxicol Appl Pharmacol. 2020;15(395):114976. doi:10.1016/j.taap.2020.114976