Background and importance Adalimumab is an anti-TNFα monoclonal antibody used in inflammatory bowel disease (IBD). Its efficacy can benefit from therapeutic drug monitoring (TDM). However, because there are several population pharmacokinetic models (PopPK) published, it is necessary to perform an evaluation of these models in the target population before being used in clinical practice.

Aim and objectives To evaluate the predictive performance and adequacy of four PopPK of adalimumab in adult patients diagnosed with IBD, using TDM in a clinical setting.

Material and methods A retrospective observational study (2014–2018) was conducted. Inclusion criteria were adult patients with IBD treated with adalimumab, with at least one trough concentration (TC). Four different PopPK were evaluated: Mod-A (FDA-2007), Mod-B Ternant-2015, Mod-C Sharma-2015 and Mod-D Berends-2018. The models were implemented in NONMEM V.7.3.

The individual and population predictions of TCs were estimated from the four PopPK models. Two datasets were created; DATASET-1 was used to evaluate the model adequacy, all patients and TCs were included, and their population predictions were compared with the observed TCs; DATASET-2 was used to assess the predictive performance and only patients with two or more TCs were included. Only the first TC of these patients was used to estimate the Bayesian estimates, and the individual predictions were compared with observed TCs.

To validate these models, bias and precision of estimated concentrations were calculated as the mean predictive error and the mean square predictive error in the population, respectively.

Results A total of 171 patients with 245 TCs in DATASET-1 and 55 patients with 74 TCs in DATASET-2 were included; 5.85% of patients in DATASET-1 and 3.64% in DATASET-2 developed anti-adalimumab antibodies.

Conclusion and relevance Mod-B performed better both in the evaluation of adequacy (DATASET-1) and for predictive performance (DATASET-2). All four models overestimated TC although Mod-B had better bias and precision (ie, closer to zero). Implementation of this PopPK in clinical practice should be done with caution.

References and/or acknowledgements No conflict of interest.