Primary nonresponse and primary nonremission are important limitations of tumor necrosis factor (TNF) antagonists, occurring in 10% to 40% and 50% to 80% of patients with inflammatory bowel disease, respectively. The magnitude of primary nonresponse differs between phase III clinical trials and cohort studies, indicating differences, e.g., in definition, patient population or blinding. The causes of nonresponse can be attributed to the drug (pharmacokinetics, immunogenicity), the patient (genetics, disease activity), the disease (type, location, severity), and/or the treatment strategy (dosing regimen, combination therapy). Primary nonresponse has been attributed to "non-TNF-driven disease" which is an overly simplified and potentially misleading approach to the problem. Many patients with primary nonresponse could successfully be treated with dose optimization during the induction phase or switching to another TNF antagonist. Therefore, primary nonresponse is frequently not a non-TNF-driven disease. Recent studies from rheumatoid arthritis and preliminary data from inflammatory bowel disease evaluating therapeutic drug monitoring have suggested that early measurement of drug and anti-drug antibody concentrations could help to define primary nonresponse and rationalize patient management of this problem. Moreover, a modeling approach including pharmacological parameters and patient-related covariants could potentially be predictive for response to the treatment. We describe an overview of this evolution in thinking, underpinned by previous findings, and assess the potential role of early measurement of drug and antidrug antibody concentrations in the definition and management of primary nonresponse.