This article reflects on methodological limitations for interpretation of relative immunogenicity of biosimilar and reference therapeutic monoclonal antibodies, in emphasizing the relevance of correlation of bioanalytical signals with appropriate clinical end points, and the possible need for post-marketing observational studies to indicate the impact of detected differences in anti-drug antibody incidence and magnitude on sustainability of treatment benefit. Given the current uncertainty regarding the longer term clinical impact of undesirable immunogenicity for reference products, there can be no predefined margin for an acceptable difference based on incidence and magnitude of detected anti-drug antibodies. Any detected differences should be assessed in relation to clinical parameters; and the designation of biosimilarity made with reference to the similarity demonstrated in the directly comparative quality, nonclinical and clinical evaluations. Application of this 'totality of evidence' approach is illustrated for infliximab and adalimumab.