The approval of biosimilars in the EU follows a comprehensive scientific assessment based on stringent regulatory standards. While the initial approach to biosimilars was understandably cautious and conservative in that uncharted territory to protect patients' safety, the analytical and scientific progress and accumulated experience with biosimilars continues to reshape regulatory requirements, generally leading to a reduced burden of clinical trials. This trend is expected to continue, for example, by increasingly employing pharmacodynamic endpoints and biomarkers, but much work remains to make this happen, especially for complex molecules with several or unknown mechanisms of action. We reviewed the available guidance and European Public Assessment Reports (EPARs) of biosimilars approved in the EU via the centralised procedure. This review focuses on the nature and extent of clinical confirmation of biosimilarity considered necessary in addition to analytical and functional data. Cases with conflicting results from different parts of the comparability exercise are discussed, with the aim of identifying whether certain elements of the comparability exercise are more important than others in determining biosimilarity. Taken together, analytical and functional comparison is the foundation of any biosimilar development. In addition, pharmacokinetic similarity is an indispensable prerequisite for any biosimilar approval, so careful planning on behalf of the applicant is mandated to avoid potential failure of such studies, for example, because of large interindividual variability, underpowered trial designs or other methodological causes. Comparative pharmacokinetic studies are a basic requirement for biosimilar development and are usually more sensitive than clinical efficacy trials when detecting potential product-related differences. This may explain why a demonstration of equivalent efficacy could not overrule a finding of dissimilar pharmacokinetic profiles in two cases of biosimilar pegfilgrastim. However, the outcome of efficacy trials depends not only on drug exposure but also on proper pharmacological action of the biological substance in vivo. Therefore, the objectives of both types of studies differ. Efficacy trials should usually be designed as equivalence trials to ensure that the efficacy of the biosimilar is neither decreased nor increased compared with the reference product. However, some remaining uncertainty regarding potentially increased efficacy of the biosimilar may be acceptable in exceptional cases, provided that the data from other parts of the comparability exercise clearly support a conclusion of biosimilarity and safety is assured. In contrast, uncertainties regarding potentially inferior efficacy of the biosimilar may not be acceptable at all. We conclude that the EU biosimilar regulatory framework is robust and able to adapt to advancing knowledge and experience and to strike a balance between regulatory standards, patient safety and feasibility of biosimilar development.