Analyzing Safety of Switching From Originators to Biosimilars: A Meta-Analysis of 21 Trials

The authors of a meta-analysis published in PLOS ONE found that serious adverse events, deaths, and treatment discontinuations observed in patients who switched between a biosimilar and reference biologic were similar to those in patients who did not switch. Additionally, adverse events linked to immunogenicity showed no significant differences between switched and non-switched patients across various biologics.

Biosimilars are becoming more available over time but, according to the investigators, some concerns persist among health care providers and patients about switching a patient whose condition is stable from a reference product to a biosimilar. Since previous reviews of safety associated with switching between an originator biologic and biosimilar have been “primarily descriptive without the use of formal statistical methods,” they conducted a meta-analysis of safety indicators.

The analysis included randomized controlled trials (RCTs) and extensions or RCTs that included 1 or more switch treatment periods (STPs) to or from a biosimilar and its reference product and a no switch control that were publicly available from the FDA, plus peer-reviewed publications that were not included in FDA reviews. Overall, 44 STPs met the eligibility criteria, from 31 unique studies on 21 biosimilars. Of these, 28 STPs included 1 switch and 16 included multiple switches. The biosimilars studied referenced 8 originator biologics: adalimumab (n = 19 STPs), epoetin-alfa (n = 2), etanercept (n = 4), filgrastim (n = 1), infliximab (n = 10), insulin-glargine (n = 1), rituximab (n = 6), and trastuzumab (n = 1). A total of 5252 patients who underwent at least 1 switch and 5770 who were not switched from the reference product were included.

Serious adverse events, deaths, treatment discontinuation was not significantly different between switch and no-switch groups

The difference in risk for serious adverse events (SAEs) between switch and no-switch groups was not significant; 436 participants in switch groups and 433 in no-switch groups experienced a SAE.The overall risk difference was –0.00 (95% CI, –0.01 to 0.01). There were 21 deaths among patients in switch groups and 23 in no-switch groups, for an overall risk difference of –0.00 (95% CI, –0.00 to 0.00). One-hundred forty-two patients discontinued treatment in switch groups compared with 160 in no-switch groups, for an overall risk difference of –0.00 (95% CI, –0.01 to 0.00).

The authors noted that it was not “scientifically appropriate” to compare antidrug antibody (ADA) levels across studies, because many patients were exposed to the reference biologic before the start of the study, and study subjects “are usually not stratified by the presence or absence of antidrug antibodies prior to a switch.” However, they were able to evaluate adverse events associated with immunogenicity, saying that these were “rare” and “not associated with switching.” When ADA data were reported at the time of an immune-related adverse event, there were no correlations between ADA status and the onset of the event.

This the first systematic review using statistical methods to evaluate the safety of switching between reference biologics and biosimilars known to the authors. They found no difference in safety profiles or immunogenicity-related events between switch and no switch groups. They added that their study addresses concerns over the safety of switching and supports reassessing the need for switching studies, “as an approved biosimilar will be analytically highly similar to its reference product.” They suggested that data-driven reports such as theirs “will facilitate efforts to streamline biosimilar development and achieve the full promise of biosimilars.”

Reference

Herndon TM, Ausin C, Brahme NN, Schrieber SJ, Luo M, Andrada FC, Kim C, Sun W, Zhou L, Grosser S, Yim S, Ricci MS. Safety outcomes when switching between biosimilars and reference biologics: A systematic review and meta-analysis. PLoS One. Published online October 3, 2023. doi:10.1371/journal.pone.0292231