For safety, the same principles of equivalence or non-inferiority trial design can be applied as for efficacy, but the endpoints and margins may differ depending on the type and severity of adverse events expected. For example, a review article by Jacobs et al. (1) states that "The safety endpoints in biosimilar trials are typically incidence rates of adverse events (AEs), serious AEs (SAEs), and AEs of special interest (AESIs). The equivalence margin for safety endpoints is usually set as a risk difference (RD) or a risk ratio (RR) between the biosimilar and the originator biologic." They also provide some examples of biosimilar trials that used different safety endpoints and margins, such as SB2 (infliximab biosimilar), SB4 (etanercept biosimilar), and SB5 (adalimumab biosimilar).

The standardisation of equivalence margins for safety is more difficult than for efficacy. For the latter, normal margins are 80-125% based on the pharmacokinetic curves which are similar among drugs. Safety requires different endpoints and statistical models that may lead to variability of equivalence margins.

(1) Jacobs I, Singh E, Sewell KL, Al-Sabbagh A, Shane LG. Patient safety and biosimilars: what we need to know. Expert Opin Drug Saf. 2018;17(1):1-10. doi: 10.1080/14740338.2018.1394459.