External validation of population pharmacokinetic models of adalimumab in adult patients with inflammatory bowel disease: Towards model-informed precision dosing

External validation of population pharmacokinetic models of adalimumab in adult patients with inflammatory bowel disease: Towards model-informed precision dosing | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article External validation of population pharmacokinetic models of adalimumab in adult patients with inflammatory bowel disease: Towards model-informed precision dosing Irene Aguilo-Lafarga, Tonet Serés-Noriega, Vicente Gimeno Ballester, and 6 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8670135/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 10 You are reading this latest preprint version Abstract Inflammatory bowel disease (IBD) has variability in the pharmacokinetics of adalimumab, which may predispose patients to subtherapeutic concentrations, therapeutic failure, and/or drug immunogenicity. Population pharmacokinetic (popPK) models estimate individual pharmacokinetic parameters and allow to personalize therapeutic regimens through Model-Informed Precision Dosing (MIPD). This retrospective, longitudinal study evaluated the predictive performance of six population pharmacokinetic models of adalimumab in adult patients with IBD treated at a tertiary hospital. The adequacy and prediction of the models were externally validated using visual analysis of the goodness-of-fit (GOF) plots, prediction-corrected visual predicted checks (pcVPC), analysis of residuals, and statistical metrics such as the Akaike Information Criteria. Bias and prediction were also calculated using mean prediction error, mean absolute percentage error and root mean square error. Bootstrap resampling was applied for statistical comparisons. Pharmacokinetic parameters were estimated using Bayesian methods and compared to theoretical values. A total of 201 subjects were included, 88% with Crohn’s disease, mean age 48.6 (± 16.2) years, 44.8% women. Overall, Berends and Vande models demonstrated the best predictive performance in the majority of comparative analyses: higher coverage in pcVPC and correlation between observed and predicted values, lower bias and precision values, the lowest AIC, and a homogeneous distribution of residuals. However, both models overestimated adalimumab population clearance. These findings support the application of the Berends and Vande models in clinical MIPD strategies. Therefore, pending broader evidence in different populations, driving MIPD adoption is critical to optimize adalimumab regimens and maximize sustained clinical outcomes. Inflammatory Bowel Diseases Adalimumab Pharmacokinetics Precision Medicine Population Pharmacokinetics Figures Figure 1 Figure 2 1. Introduction Inflammatory bowel disease (IBD), which includes Crohn’s disease (CD) and ulcerative colitis (UC), is a chronic disease characterized by significant morbidity and impaired quality of life [ 1 ]. Biologic therapies, in particular adalimumab (ADA), the most used fully human monoclonal antibody directed against tumor necrosis factor alpha (anti-TNFα), have substantially advanced the management of moderate-to-severe CD and UC by achieving and sustaining remission [ 2 , 3 ]. Regarding to ADA, the most common treatment strategy consists of administering in a standardized dosage regimen, which does not take into account the differences between individuals and the variability in drug metabolism and elimination [ 1 ]. Notably, the pharmacokinetics of ADA in patients with IBD are highly variable so standard dosing may lead to suboptimal plasma levels [ 4 ]. In turn, low ADA concentrations have been associated with increased risk of treatment failure and development of anti-adalimumab antibodies (AAA) [ 5 – 9 ]. These limitations have led to empirical adjustments in dose and/or frequency of administration in clinical practice to optimize clinical response and to ensure the effectiveness and safety of treatment [ 10 , 11 ]. Accordingly, population pharmacokinetic (popPK) models enable the quantification of pharmacokinetic variability among patients treated with ADA, as they incorporate individual variables such as gender, weight, C-reactive protein (CRP), faecal calprotectin (FCP), and albumin, among others. Model-Informed Precision Dosing (MIPD) based on popPK models is a tool that allows treatment optimization using Bayesian statistical estimates that predict ADA distribution and elimination for each patient and help adjust doses and/or dosing intervals [ 12 ]. In this way, the MIPD seeks to individualize treatment dosages to promote the rational use of the drug, maximizing its safety and effectiveness [ 13 , 14 ]. Its use through therapeutic drug monitoring (TDM) and analysis of AAA in serum has become a recognized clinical strategy [ 15 , 16 ]. Although evidence remains scarce, many studies point to its usefulness in reducing therapeutic failure and towards early optimization of ADA dosing [ 14 , 17 ]. However, given that popPK models are developed using distinct patient cohorts, their predictive performance may vary across different clinical settings. In addition, popPK models have generally been conducted in small cohorts, in different geographical contexts and often excluding patients with UC, making it essential to conduct external validation studies in larger samples that are representative of actual clinical practice. In this context, the aim of this study was to compare the main adult IBD patients popPK models of adalimumab in order to analyze their predictive performance in a representative and significant sample of a tertiary hospital. As an exploratory analyses, we also assessed the associations between HLA-DQA1*05 and ADA concentrations. 2. Methodology 2.1 Study design and participants This was a retrospective and longitudinal study, performed at a tertiary care hospital, that consecutively recruited all IBD patients treated with ADA between January/2022 to October/2024 for whom at least two serum measurements of ADA concentration are available. The reporting of this study was conducted in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement. Demographic, clinical and pharmacokinetic data were collected from medical records. Patients for whom the necessary information to perform the pharmacokinetic analysis in the different study models was not available were excluded. ADA treatment consisted in an induction phase that was established as 160mg and 80mg at weeks 0 and 2 respectively. Following this phase, as a maintenance phase, patients were treated with 40mg of adalimumab every other week. If clinically indicated and under medical prescription, treatment intensification was allowed. 2.2 Laboratory analysis Each sample was collected before the corresponding dose of ADA during induction and/or maintenance phase (trough levels). ADA serum concentrations were measured using an enzyme-linked immunosorbent assay (Promonitor®-ADA, Grifols Diagnostic Solutions Inc., California, US). The limits of quantification were: 0.25 and 12.0µg/ml. For samples above the upper limit, additional dilution was required. AAA determination was performed using a drug-sensitive technique, consequently, AAA detection was limited to samples lacking detectable ADA. Therefore, the presence of AAA was only evaluated upon request by the clinician in patients with therapeutic failure and a high suspicion of immunogenicity, defined as unexplained decline in adalimumab serum concentrations. ADA concentrations ≥ 5 µg/mL and ≥ 6.4 µg/mL at induction phase were considered within the therapeutic range for CD and UC, respectively. Therapeutic levels during the maintenance phase were defined as ADA concentrations > 10–12 µg/mL, associated with endoscopic healing and clinical remission [ 16 ]. Screening of HLA-DQA1*05 genotype was realized with the polymerase chain reaction method, using the CeliacStrip® test (Operon S.A., Zaragoza, Spain). Other laboratory variables necessary for clinical follow-up were also collected, such as CRP, FCP, serum albumin, hemoglobin and erythrocyte sedimentation rate. 2.3 Validation of models Six adult popPK models were selected: Berends, FDA, Marquez, Sanchez, Ternant and Vande, respectively [ 18 – 23 ]. All of them describe ADA pharmacokinetics using a one-compartment structure with first-order absorption in adult patients. ADA popPK models and equations are shown in Supplementary Tables 1 and 2 . The predictive performance of every model was evaluated by visual analysis of the goodness-of-fit (GOF) plots, and prediction-corrected visual predicted checks (pcVPC). In GOF plots a good predictive performance was defined as a close alinement of the data points along the identity line. The correlation in GOF plots is presented solely as a descriptive measure. Coverage, defined as the proportion of observed ADA serum concentrations falling within the simulated 90% prediction intervals, was computed for every model in line with the pcVPC. A good predictive performance was considered to occur when the empirical percentiles were contained within these intervals, and higher coverage values indicated better agreement between observed and simulated data. We calculated the distribution of the Empirical Bayesian estimate (EBEs) of the pharmacokinetic parameters for each of the popPK models. This data was compared with the theoretical distribution of these pharmacokinetic parameters according to every popPK model. Bias and precision of the models were calculated through mean prediction error (MPE), mean absolute prediction error (MAPE) and root mean square error (RMSE) based on the following formulas: in which Ỹ represents the population predicted ADA concentration, while Y represents the observed ADA concentration and n is the total of observations analyzed. $$\:MPE=\frac{\sum\:(Ỹ-Y)}{n}\:\:\:\:\:\:\:\:\:\:\:\:\:MAPE=\left|\frac{\sum\:(Y\:-Ỹ)}{Ỹ}\right|x\frac{100}{n}\:RMSE=\sqrt{\frac{{\sum\:(Ỹ-Y)}^{2}}{n}}$$ This study used a bootstrap resampling approach to compare the statistical behavior of the three metrics employed. Baseline metrics were first calculated using paired observations of actual and predicted values; then 1000 bootstrap samples were generated by random resampling with replacement. Each bootstrap sample was used to recalculate the metrics, thereby obtaining empirical distributions from which means, standard deviations, and 95% confidence intervals were estimated. The statistical comparisons were adjusted for multiple comparisons using Bonferroni correction to avoid spurious statistical interpretations of the results. Further, the structural models were compared using Akaike Information Criterion (AIC), with lower AIC values indicating better model fit. Such a comparison of both the metric distributions and AIC values allowed for a robust assessment of the variability and bias, ensuring a reproducible framework for performance evaluation. 2.4 Statistical analysis The Shapiro-Wilk test and P-P plots were used to assess the distribution of variables. Comparisons between variables were made using Chi-square statistical tests for qualitative variables; Student’s t-test for comparison between means if normally distributed variables; Mann-Whitney test, Wilcoxon test and Kruskal-Wallis test were used for comparisons between groups with non-parametric distribution. The correlation between observed and individual predicted values was measured using a Pearson’s correlation analysis. A p-value of < 0.05 was considered significant. Analyses were carried out with STATA/MP® v.17 and R-Studio® v.2024.12.0. Simulation and estimation of pharmacokinetic models were performed using MonolixSuite™ v.2024.R1. 3. Results 3.1 Sample characteristics A total of 201 patients were included in the study, of whom 177 (88.1%) had Crohn’s disease (CD). Females represented 44.8%, the mean basal age was 48.6±16.2 years and mean body mass index (BMI) was 25.3±5.2 kg/m 2 , with no significant difference between CD and UC participants ( p > 0.126). Years of disease progression were significantly longer in UC patients (14.8; IQR: [4.7-19.5]) than in CD (5.8; IQR [2.4-14.2]; p = 0.018) at the start of pharmacokinetic monitorization. The majority of patients (72.6%) received treatment with immunomodulators. Although these were mainly prescribed prior to receiving adalimumab, 27.9% of the subjects received them concomitantly with ADA. The most commonly used immunomodulator was azathioprine (66.2%). Methotrexate was only used among CD patients (14.4%). Prior anti-TNFα therapy had been administered to 11.9% of the cohort, predominantly infliximab. The rest of data are shown in Table 1. Data according to Montreal classification is shown in Supplementary Table 3. A total of 709 ADA serum samples were analyzed, mostly during maintenance phase (74.8%). The average amount of ADA serum samples per subject was 3 (IQR [2-4]). The overall mean ADA serum concentration was 13.8±4.5 mg/L at week 2, 13.4±6.2 mg/L at week 6 and 12.3±7.1 mg/L during maintenance phase. In both the induction and maintenance phases, no differences in ADA concentrations were observed between EC and UC. However, a significantly higher proportion of patients with UC had ADA concentrations within the therapeutic target range compared to those with CD (80.5% vs. 64.8%; p=0.004). This is probably why, mainly in the CD group the dosage regimen was increased to a weekly basis in 14.4% of all patients, while 6.2% of the subjects had their maintenance dose increased to 80 mg. No differences were observed in ADA concentrations in patients with perianal disease (p=0.12). Overall, AAA were detected in 5.5% of the samples, all of which occurred in CD patients. An exploratory analysis was conducted to compare groups with and without HLA-DQA1*05. The allele was present in 38% of the evaluated patients, without significant differences between CD and UC groups ( p = 1.000). ADA concentrations within the therapeutic range were achieved by 67% of individuals with HLA-DQA1*05 and by 70.7% of patients without the allele (p=0.215). Neither were there any significant differences in ADA serum concentrations (12.8±6.9 mg/L and 12.4±5.5 mg/L for patients without and with the allele, respectively). Only 6 of the patients with AAA were tested for the presence of HLA-DQA1*05. No association was observed between HLA-DQA1*05 and AAA formation (1.07% and 1.19% for the absent and present allele, respectively, p=0.631). 3.2 Validation of models As shown in Figure 1 , the correlation between predicted and observed adalimumab concentrations varied substantially among the compared popPK models. Of these, the Vande and Berends models displayed the strongest correlations, at r=0.83 and r=0.77 (p<0.001 for both), respectively, with data points in close proximity to the line of identity. In contrast, the Sanchez and Ternant’s models exhibited weak to moderate correlations (r=0.45 and r=0.37; p<0.001), representing the poorest correlations among the models evaluated and the largest scatter around the line of identity. These findings were further confirmed by the predictive check with pcVPC plots presented in Figure 2 . The model of Berends showed the best performance: more than 90% of the observed data fell within the 90% prediction interval. The models of Vande and FDA also performed well, with more than 80% of observations covered. In contrast, for all the other models, a considerable amount of data fell outside of the prediction intervals, reflecting systematic bias and lower precision. More importantly, all models provided a negative prediction of the central tendency; thus, the simulated medians were consistently lower than the respective observed median values. As shown in Table 2 , ADA clearance was generally overestimated in most models, except for Sánchez and Ternant’s models, whose population and individual estimates were closely aligned. These observations are also consistent with the performance metrics outlined in Table 3 and Supplementary Figures 1-3 . The MPE showed significant differences for most models at p < 0.001, except between Sanchez and Ternant’s models, which were comparably poor in their predictive abilities (p = 0.112). Conversely, the models of Berends and Vande had respective MPE values of 0.58 and –0.43 µg/mL, which reflected minimal bias and a strong agreement between the predicted and observed values. Interestingly, in the Sanchez’s model, the MAPE was the lowest at 1.37% (95%CI: 1.32-6.71), indicating that it was relatively more accurate in its predictions. However, its MAPE values were more spread out indicating more variability in the individual estimates. The Marquez’s model had a relatively low MAPE of 1.86% (95%CI: 0.45-3.28), consistent with its modest MPE, indicating an acceptable model fit. Finally, the model developed by FDA had the highest MAPE 7.0% (2.35-11.65) and was thus the least reliable. In terms of RMSE comparison between models, the Vande and Berends models were again the best (3.85 µg/mL (95%CI: 3.55-4.14) and 4.68 µg/mL (95%CI: 4.19-5.16) µg/mL, respectively), reflecting the highest predictive precision. The poorest performances regarding the RMSE of estimates included Sanchez and Ternant models 7.51µg/mL (95%CI: 6.78-8.24) and 7.84µg/mL (95%CI: 6.99-8.71), respectively. Moreover, the Berends and Vande models also demonstrated the lowest AIC values, reinforcing their superior model fit. Residual distribution in both the Berends and Vande’s models resembled normality, was centered at zero, and showed homogeneous dispersion, which suggests good overall fit and absence of systematic bias. Graphs of residuals versus individual prediction and versus time showed that in both models, random dispersion without structured patterns was maintained throughout, supporting consistency of the model across a wide range of concentrations and follow-up time. By contrast, the models from the FDA, Marquez, Sanchez, and Ternant had greater dispersion, were heteroscedastic, with residuals further away from zero with noticeable clusters in some time periods or prediction levels. This analysis suggests lower predictive capacity in more variable clinical scenarios. Figures are shown in Supplementary Figure 4 . 4. Discussion In the current study, we performed an extensive external validation of six published popPK models of ADA in a cohort of 201 IBD patients. We compared their predictive performance to identify which model best describes ADA pharmacokinetics in this population. Our results consistently confirm the models of Berends and Vande as most reliable in predicting adalimumab pharmacokinetics within this population. Given that they appear superior across multiple metrics for evaluation, their superiority is indicative of the best balance between bias, precision, and overall predictive accuracy. The superiority of Berends [ 18 ] and Vande [ 23 ] models was evident in highest closeness to the line of identity in GOF plots and favorable pcVPC plots, with more than 80% of observed values falling within the 90% prediction interval. These findings indicate a good match between model-derived predictions and real-world ADA concentrations. In addition, both models showed low MPE, low RMSE, and homogeneously distributed residuals centered around zero, key indicators of low bias and high precision [ 24 ]. However, the simulated median tended to underestimate the observed median concentrations, which is partly justified by the overestimation of ADA population clearance. Nevertheless, this bias is small and does not compromise the utility of these popPK models for MIPD, as Bayesian updating corrects for population-level discrepancies when patient-specific data are incorporated. The suitability of the Berends model is consistent with a study conducted by Marquez-Megias S, et al [ 25 ]. Although the population studied differs from ours, as it evaluates patients with biomarkers indicative of a greater IBD activity (with lower serum albumin, higher CRP and FCP), the suitability of Berends [ 18 ] model is probably due to the selected model's covariates. As shown in Supplementary Table 2 , Berends model incorporates the dosing regimen (every other week vs. every week), which is particularly relevant given that in our population 23.4% of subjects receive every week regimen. This covariate allows the effect of the regimen to be reflected on clearance, improving the model's ability to reproduce the concentrations observed in typical clinical situations of treatment optimization. In addition, Vande's model incorporates the lean body weight (LBW) as a main covariate. We believe that LBW is particularly appropriate for our cohort since it is normally distributed and do not include extreme values, unlike other variables proposed by other models (such as FCP or serum albumin etc.). Consequently, the relationship between LBW and ADA clearance is more consistent and less affected by outliers. This contributes to LBW being a robust covariate and helps to improve the predictive capacity of the model in comparison to other variables. In contrast, the models by Sanchez [ 21 ] and Ternant [ 22 ], despite estimating clearance values similar to those in our cohort, demonstrated poor predictive performance. Although in 11.9% of our patients ADA immunogenicity was suspected, tests to detect the presence of AAA were only performed in 11 of them, resulting in a low prevalence of positive AAA in our cohort (5.5%). The large impact of AAA on ADA clearance may have contributed to the robustness of Berends [ 18 ] and Vande [ 23 ] models, but Ternant model [ 22 ], by proposing only this variable as a predictor of ADA clearance, results in most patients in our study having a fixed clearance, leading to no significant differences between theoretical and estimated clearance, but without the ability to reproduce the actual variability between patients. This lack of explained variability is reflected in the poor alignment of individual predictions and insufficient pcVPC. It is likely that if the proportion of AAAs evaluated were higher in our subjects, Ternant model would predict better, as is the case in other studies [ 25 ]. Contrary, in Sanchez model [ 21 ], this discrepancy is probably due to its lack of parsimony. The structural complexity and high number of covariates may have led to overfitting in the original population, limiting its extrapolation capacity and explaining the poor individual fit and low coverage of the pcVPC in our study. The FDA model [ 19 ] also performed reasonably well. However, its lack of covariates explaining clearance, limits its ability to capture interindividual variability in ADA clearance. In contrast, the model developed by Marquez [ 20 ] was based on patients with a presumably more severe disease activity, characterized by lower albumin, higher FCP and CRP levels. These differences in baseline characteristics may be partly responsible for the lower accuracy observed when applying these models to our dataset. In our cohort, the presence of the HLA-DQA1*05 allele was not associated with differences in serum ADA concentrations, in the proportion of patients who reached therapeutic levels, or in the formation of AAA. These results differ from previous studies that have associated the allele with an increased immunogenicity against anti-TNFα and subtherapeutic concentrations of anti-TNFα, but no greater therapeutic failure [ 26 – 28 ]. Our results could be explained by the small sample size, especially in the AAA group, and by the high proportion of patients with empirically intensified regimens (33.8%). This study presents several strengths, including the relatively large cohort, a high number of validated samples per subject, and its conduct in real-world clinical practice involving patients with IBD represents an unselected sample of all patients who initiated ADA at our hospital. All popPK models evaluated were developed in patients with IBD, in contrast to most previous studies that have attempted to extrapolate popPK models developed for other autoimmune diseases, such as psoriasis or rheumatoid arthritis, to populations with IBD [ 18 , 29 ]. In addition, the methodology used is consistent and rigorous, with multiple validation metrics employed. This facilitates the extrapolation of the findings to other contexts, thereby enhancing the external validity of the study. Nonetheless, some limitations must be acknowledged. First, the retrospective design may introduce biases related to sampling time and in data collection. Second, we lacked data on baseline disease activity, so related secondary variables were collected. We consider the external validation approach used in the study an appropriate first step toward implementing MIPD in clinical practice, facilitating individualized ADA dosing in patients with IBD. This is particularly relevant for CD patients, who in our study exhibited a higher frequency of subtherapeutic ADA levels compared to those with UC. Future research should focus on conducting prospective studies, exploring how other covariates, such as HLA-DQA1*05, influence ADA clearance, and developing popPK models with sufficient external validity to allow their extrapolation to different patients and clinical contexts. In conclusion, the superior predictive performance of the Berends and Vande models underscores their pivotal role in modernizing adalimumab therapy. Our results provide a solid foundation for implementing MIPD, moving beyond empirical adjustments. Therefore, pending further evidence from diverse populations, it is crucial to prioritize the adoption of these models to maximize therapeutic success and minimize bias in clinical decision-making. Declarations Funding: No funding was received to assist with the preparation of this manuscript. Competing interests: The authors declare no conflicts of interest. Data Availability Statement: The datasets generated during the current study are available from the corresponding author on reasonable request. Ethics approval: The study was approved by the Clinical Research Ethics Committee of Autonomous Community of Aragon (CEICA) with the following registration number: EOM24/034. The procedures established in this research are designed following the principles of good clinical practice and the declaration of Helsinki. Informed consent: All participants provided written informed consent before enrollment in the study. Author Contributions: All authors contributed to the study conception and design. Data collection and analysis were performed by Irene Aguilo-Lafarga, Tonet Serés-Noriega, Vicente Gimeno Ballester and Pilar Antonia Corsino Roche. The first draft of the manuscript was written by Irene Aguilo-Lafarga. Tonet Serés-Noriega, Vicente Gimeno Ballester, Pilar Antonia Corsino Roche, Raquel Vicente Lidon, Eva Maria Sierra Moros, Sonia Gallego, Santiago Garcia Lopez and Maria Reyes Abad Sazatornil reviewed the relevant literature and provided a critical review of the manuscript. References Moran GW, Gordon M, Sinopolou V, et al (2025) British Society of Gastroenterology guidelines on inflammatory bowel disease in adults: 2025. Gut 74:s1–s101. https://doi.org/10.1136/gutjnl-2024-334395 Gordon H, Minozzi S, Kopylov U, et al (2024) ECCO Guidelines on Therapeutics in Crohn’s Disease: Medical Treatment. J Crohns Colitis 18:1531–1555. https://doi.org/10.1093/ecco-jcc/jjae091 Raine T, Bonovas S, Burisch J, et al (2022) ECCO Guidelines on Therapeutics in Ulcerative Colitis: Medical Treatment. J Crohns Colitis 16:2–17. https://doi.org/10.1093/ecco-jcc/jjab178 Vande Casteele N, Gils A. (2015) Pharmacokinetics of anti-TNF monoclonal antibodies in inflammatory bowel disease: Adding value to current practice. J Clin Pharmacol. 55:S39–50. https://doi.org/10.1002/jcph.374 Roblin X, Marotte H, Rinaudo M, et al (2014) Association between pharmacokinetics of adalimumab and mucosal healing in patients with inflammatory bowel diseases. Clin Gastroenterol Hepatol 12:80-84.e2. https://doi.org/10.1016/j.cgh.2013.07.010 Rueda García JL, Suárez-Ferrer C, Amiama Roig C, et al (2025) Association of early therapeutic drug monitoring of adalimumab with biologic remission and drug survival in Crohn’s Disease. Therap Adv Gastroenterol 18:17562848251324226. https://doi.org/10.1177/17562848251324226 Yanai H, Lichtenstein L, Assa A, et al (2015) Levels of drug and antidrug antibodies are associated with outcome of interventions after loss of response to infliximab or adalimumab. Clin Gastroenterol Hepatol 13:522-530.e2. https://doi.org/10.1016/j.cgh.2014.07.029 Ungar B, Levy I, Yavne Y, et al. (2016) Optimizing Anti-TNF-α Therapy: Serum Levels of Infliximab and Adalimumab Are Associated With Mucosal Healing in Patients With Inflammatory Bowel Diseases. Clin Gastroenterol Hepatol. 14:550-557.e2. https://doi.org/10.1016/j.cgh.2015.10.025 Kennedy NA, Heap GA, Green HD, et al (2019) Predictors of anti-TNF treatment failure in anti-TNF-naive patients with active luminal Crohn’s disease: a prospective, multicentre, cohort study. Lancet Gastroenterol Hepatol 4:341–353. https://doi.org/10.1016/S2468-1253(19)30012-3 Ma C, Huang V, Fedorak DK, et al (2014) Adalimumab dose escalation is effective for managing secondary loss of response in Crohn’s disease. Aliment Pharmacol Ther 40:1044–1055. https://doi.org/10.1111/apt.12940 Sandborn WJ, Colombel J-F, Schreiber S, et al (2011) Dosage adjustment during long-term adalimumab treatment for Crohn’s disease: clinical efficacy and pharmacoeconomics. Inflamm Bowel Dis 17:141–151. https://doi.org/10.1002/ibd.21328 Minichmayr IK, Dreesen E, Centanni M, et al (2024) Model-informed precision dosing: State of the art and future perspectives. Adv Drug Deliv Rev 215:115421. https://doi.org/10.1016/j.addr.2024.115421 Papamichael K, Stocco G, Ruiz Del Agua A (2023) Challenges in Therapeutic Drug Monitoring: Optimizing Biological Treatments in Patients With Inflammatory Bowel Disease and Other Immune-Mediated Inflammatory Diseases. Ther Drug Monit 45:579–590. https://doi.org/10.1097/FTD.0000000000001095 Hanauer SB, Torres EA, Aragon-Han P, et al (2025) The Clinical Utility of Precision-Guided Dosing for Adalimumab Therapy Optimization in Inflammatory Bowel Disease: A Clinical Experience Program. Pharmaceutics 17:428. https://doi.org/10.3390/pharmaceutics17040428 Papamichael K, Cheifetz AS, Melmed GY, et al (2019) Appropriate Therapeutic Drug Monitoring of Biologic Agents for Patients With Inflammatory Bowel Diseases. Clin Gastroenterol Hepatol 17:1655-1668.e3. https://doi.org/10.1016/j.cgh.2019.03.037 Rodríguez-Moranta F, Argüelles-Arias F, Hinojosa Del Val J, et al (2024) Therapeutic drug monitoring in inflammatory bowel diseases. Position statement of the Spanish Working Group on Crohn’s Disease and Ulcerative Colitis. Gastroenterol Hepatol 47:522–552. https://doi.org/10.1016/j.gastrohep.2024.01.007 Papamichael K, Juncadella A, Wong D, et al (2019) Proactive Therapeutic Drug Monitoring of Adalimumab Is Associated With Better Long-term Outcomes Compared With Standard of Care in Patients With Inflammatory Bowel Disease. J Crohns Colitis 13:976–981. https://doi.org/10.1093/ecco-jcc/jjz018 Berends SE, Mould DR, Strik AS, et al. (2018) Explaining Interpatient Variability in Adalimumab Pharmacokinetics in Patients With Crohn’s Disease. Ther Drug Monit.40:202–11. https://doi.org/10.1097/ftd.0000000000000494 Food and Drug Administration (FDA) (2008). Humira: Clinical Pharmacology and Biopharmaceutics Review(s). FDA. Available online: https://www.accessdata.fda.gov/drugsatfda_docs/bla/2007/125057_S0089.pdf. Accessed 28 Nov 2025. Marquez-Megias S, Nalda-Molina R, Más-Serrano P, Ramon-Lopez A. (2023) Population Pharmacokinetic Model of Adalimumab Based on Prior Information Using Real World Data. Biomedicines. 11:2822. https://doi.org/10.3390/biomedicines11102822 Sánchez-Hernández JG, Pérez-Blanco JS, Rebollo N, et al (2020) Biomarkers of disease activity and other factors as predictors of adalimumab pharmacokinetics in inflammatory bowel disease. Eur J Pharm Sci 150:105369. https://doi.org/10.1016/j.ejps.2020.105369 Ternant D, Karmiris K, Vermeire S, et al. (2015) Pharmacokinetics of adalimumab in Crohn’s disease. Eur J Clin Pharmacol. 71:1155–7. https://doi.org/10.1007/s00228-015-1892-1 Vande Casteele N, Baert F, Bian S, et al (2019) Subcutaneous Absorption Contributes to Observed Interindividual Variability in Adalimumab Serum Concentrations in Crohn’s Disease: A Prospective Multicentre Study. J Crohns Colitis 13:1248–1256. https://doi.org/10.1093/ecco-jcc/jjz050 Sherwin CMT, Kiang TKL, Spigarelli MG, Ensom MHH (2012) Fundamentals of population pharmacokinetic modelling: validation methods. Clin Pharmacokinet 51:573–590. https://doi.org/10.1007/BF03261932 Marquez-Megias S, Ramon-Lopez A, Más-Serrano P, et al (2021) Evaluation of the Predictive Performance of Population Pharmacokinetic Models of Adalimumab in Patients with Inflammatory Bowel Disease. Pharmaceutics 13:1244. https://doi.org/10.3390/pharmaceutics13081244 Navajas Hernández P, Mouhtar El Halabi S, González Parra AC, et al (2024) Carriage of the HLA-DQA1⋆05 haplotype is associated with a higher risk of infratherapeutic drug concentration and higher immunogenicity in patients undergoing treatment with anti-TNF for inflammatory bowel disease. Therap Adv Gastroenterol 17:17562848241278145. https://doi.org/10.1177/17562848241278145 Pascual-Oliver A, Casas-Deza D, Cuarán C, et al (2024) HLA-DQA1*05 Was Not Associated With Primary Nonresponse or Loss of Response to First Anti-TNF in Real-World Inflammatory Bowel Disease Patients. Inflamm Bowel Dis 30:922–929. https://doi.org/10.1093/ibd/izad130 Reppell M, Zheng X, Dreher I, et al (2025) HLA-DQA1*05 Associates With Anti-Tumor Necrosis Factor Immunogenicity and Low Adalimumab Trough Concentrations in Inflammatory Bowel Disease Patients From the SERENE Ulcerative Colitis and Crohn’s Disease Studies. J Crohns Colitis 19:jjae129. https://doi.org/10.1093/ecco-jcc/jjae129 Martínez-Romero GJ, Alvariño A, Hinojosa E, et al (2019) Validation of a population pharmacokinetic model of adalimumab in a cohort of patients with inflammatory bowel disease. Rev Esp Enferm Dig 111:431–436. https://doi.org/10.17235/reed.2019.5600/2018 Tables Table 1. Baseline clinical characteristics of the sample according to the type of inflammatory bowel disease Global (n=201) Crohn’s disease (n=177) Ulcerative colitis (n=24) p -value Sex (female) 88 (44.8) 74 (41.8) 14 (58.3) 0.126 Basal age (years) 48.6 (±16.2) 47.9 (±16.0) 53.8 (±17.8) 0.951 IBD duration (years) 6.1 (2.4-15.3) 5.8 (2.4-14.2) 14.8 (4.7-19.5) 0.018 Weight (kg) 72.2 (±17.1) 71.9 (±16.9) 70.9 (±13.4) 0.383 Height (cm) 168.5 (±9.4) 168.6 (±9.3) 167.8 (±10.3) 0.354 BMI (kg/m 2 ) 25.3 (±5.2) 25.2 (±5.2) 25.0 (±3.7) 0.447 LBW (kg) 49.7 (±0.5) 49.8(±0.5) 49.6(±1.4) 0.445 BSA (m 2 ) 1.82 (±0.2) 1.83 (±0.24) 1.81 (±0.22) 0.397 Albumin (g/dL) 4.2 (4.0-4.4) 4.2 (4.0-4.4) 4.2 (4.0-4.5) 0.120 CRP (mg/dL) 0.35 (0.11-0.84) 0.36 (0.11-0.84) 0.21 (0.10-0.76) 0.584 FCP (mg/kg) 107.8 (50.2-261.8) 104.0 (50.2-244.1) 244.8 (49.1-674.2) 0.060 Hemoglobin (g/dL) 13.9 (±1.6) 13.9 (±1.6) 13.8 (±1.6) 0.390 ESR (mm/h) 12 (5-22) 12 (5-21) 15(4-32) 0.664 Extraintestinal manifestations ∫ 40 (19.9) 36 (20.3) 4 (16.7) 0.791 Concomitant immunomodulator treatment 6-Mercaptopurine 13 (6.5) 12 (7.9) 1 (4.2) 1.000 Azathioprine 133 (66.2) 119 (67.2) 14 (58.3) 0.387 Methotrexate 29 (14.4) 29 (16.4) 0 (0.0) 0.029 Previous anti-TNFα treatment 24 (11.9) 21 (11.9) 3 (12.5) 1.000 Positive HLA-DQA1*05 ǂ 52 (38.0) 47 (37.9) 5 (38.5) 1.000 Previous IBD surgery 47 (23.5) 46 (26.0) 1 (4.2) 0.017 Active tobacco use 46 (22.9) 44 (24.9) 2 (8.3) 0.075 ADA serum samples Ꝭ 709 (100) 622 (87.7) 87 (12.3) - Induction Ꝭ 179 (25.2) 157 (25.2) 22 (25.3) 0.980 Maintenance Ꝭ 530 (74.8) 465 (74.8) 65 (74.7) 0.980 ADA serum concentrations (mg/L) 12.6 (±6.7) 11.6 (7.6-16.8) 12.8 (±4.9) 0.319 ADA serum concentrations in target range Ꝭ 473 (67.6) 403 (64.8) 70 (80.5) 0.004 Suspected immunogenicity ֏ 24 (11.9) 23 (11.4) 1 (4.2) 0.343 AAA positive 11 (5.5) 11 (6.2) 0 (0.0) 0.377 Data are shown as n (percentage within each column), mean (±SD) or median (Q1-Q3). p values for group comparisons are reported ADA: Adalimumab; BMI: Body mass index; BSA: Body surface area; CRP: C-reactive protein; ESR: erythrocyte sedimentation rate; FCP: faecal calprotectin; IBD: inflammatory bowel disease; LBW: lean body weight; TNFα: tumor necrosis factor α. ∫ Extraintestinal manifestations included: axial spondyloarthropathy, peripheral arthropathy, erythema nodosum, pyoderma gangrenosum, sweet's syndrome, oral ulcers, episcleritis, uveitis and primary sclerosing cholangitis. ǂ Available in n = 137; 124 patients with Crohn's disease and 13 with ulcerative colitis. Ꝭ Percentage within total adalimumab serum samples. ֏ Unexplained decline in adalimumab serum concentrations. Table 2. Comparison between theorical and Estimated Bayesian Estimates (EBEs) clearance among models Berends FDA Marquez Sanchez Ternant Vande Theorical CL/F (L/day) 0.32 (0.32-0.32) 0.30 (0.30-0.30) * 1.22 (0.98-1.66) 0.41 (0.35-0.48) 0.42 (0.42-0.42) 0.36 (0.29-0.43) EBEs CL/F (L/day) 0.25 (0.19-0.35) 0.26 (0.21-0.33) 0.36 (0.26-0.52) 0.41 (0.35-0.48) 0.42 (0.42-0.42) 0.26 (0.19-2.51) p-value <0.001 <0.001 0.05 >0.05 <0.001 Data are shown as mean (95%CI). CL/F: clearance; EBEs: Estimated Bayesian Estimates * Fixed CL/F Table 3. Bias, precision (µg/mL) and Akaike Information Criteria of analyzed population PK models MPE (µg/mL) MAPE (%) RMSE (µg/mL) AIC Berends 0.58 (0.24 to 0.93) 2.77 (0.89 to 4.67) 4.68 (4.19 to 5.16) 4470 FDA -1.18 (-1.53 to -0.84) 7.00 (2.35 to 11.65) 4.78 (4.41 to 5.16) 4729 Marquez -2.88 (-3.25 to -2.50) 1.86 (0.45 to 3.28) 5.80 (5.27 to 6.33) 5547 Sanchez -4.56 (-5.01 to -4.12) 1.37 (1.32 to 6.71) 7.51 (6.78 to 8.24) 13121 Ternant -4.76 (-5.22 to -4.30) 4.37 (1.22 to 7.52) 7.84 (6.99 to 8.71) 11923 Vande -0.43 (-0.71 to -0.14) 2.64 (0.65 to 4.63) 3.85 (3.55 to 4.14) 4700 Data are shown as mean (95%CI). AIC were truncated to whole numbers. AIC: Akaike Information Criteria; CI: confidence interval; MAPE: mean absolute percentage error; MPE: mean prediction error; RMSE: root mean square error. Additional Declarations No competing interests reported. Supplementary Files Supplementarymaterial.docx SlFig4.Analysisoftheresiduals.tiff Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 10 Mar, 2026 Reviews received at journal 23 Feb, 2026 Reviewers agreed at journal 11 Feb, 2026 Reviewers agreed at journal 09 Feb, 2026 Reviews received at journal 03 Feb, 2026 Reviewers agreed at journal 29 Jan, 2026 Reviewers invited by journal 29 Jan, 2026 Editor assigned by journal 23 Jan, 2026 Submission checks completed at journal 23 Jan, 2026 First submitted to journal 22 Jan, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8670135","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":582394682,"identity":"cf9ab55d-c6d3-4cf2-af75-b4d4b1da29e3","order_by":0,"name":"Irene Aguilo-Lafarga","email":"data:image/png;base64,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","orcid":"","institution":"Hospital Universitario Miguel Servet","correspondingAuthor":true,"prefix":"","firstName":"Irene","middleName":"","lastName":"Aguilo-Lafarga","suffix":""},{"id":582394683,"identity":"690bfeaf-ea39-45c0-8cd0-7df837292a19","order_by":1,"name":"Tonet Serés-Noriega","email":"","orcid":"","institution":"Centro Médico Milenium","correspondingAuthor":false,"prefix":"","firstName":"Tonet","middleName":"","lastName":"Serés-Noriega","suffix":""},{"id":582394684,"identity":"c0239792-841c-407e-baf9-0c06498db5da","order_by":2,"name":"Vicente Gimeno Ballester","email":"","orcid":"","institution":"Hospital Universitario Miguel Servet","correspondingAuthor":false,"prefix":"","firstName":"Vicente","middleName":"Gimeno","lastName":"Ballester","suffix":""},{"id":582394685,"identity":"cc8d3a60-a292-4abb-b406-bbea932e363f","order_by":3,"name":"Pilar Antonia Corsino Roche","email":"","orcid":"","institution":"Hospital Universitario Miguel Servet","correspondingAuthor":false,"prefix":"","firstName":"Pilar","middleName":"Antonia Corsino","lastName":"Roche","suffix":""},{"id":582394686,"identity":"7ab7ccc7-113b-49ae-9a95-9aa1d834c6c4","order_by":4,"name":"Raquel Vicente Lidon","email":"","orcid":"","institution":"Hospital Universitario Miguel Servet","correspondingAuthor":false,"prefix":"","firstName":"Raquel","middleName":"Vicente","lastName":"Lidon","suffix":""},{"id":582394687,"identity":"008da6ea-6777-400e-954f-e7ee77a02627","order_by":5,"name":"Eva Maria Sierra Moros","email":"","orcid":"","institution":"Hospital Universitario Miguel Servet","correspondingAuthor":false,"prefix":"","firstName":"Eva","middleName":"Maria Sierra","lastName":"Moros","suffix":""},{"id":582394690,"identity":"58c8341d-f0f4-4ceb-b592-ad42d338edf3","order_by":6,"name":"Sonia Gallego","email":"","orcid":"","institution":"Hospital Universitario Miguel Servet","correspondingAuthor":false,"prefix":"","firstName":"Sonia","middleName":"","lastName":"Gallego","suffix":""},{"id":582394692,"identity":"aafb795e-45fe-4c39-83b2-52be1d9802a3","order_by":7,"name":"Santiago Garcia Lopez","email":"","orcid":"","institution":"Hospital Universitario Miguel Servet","correspondingAuthor":false,"prefix":"","firstName":"Santiago","middleName":"Garcia","lastName":"Lopez","suffix":""},{"id":582394694,"identity":"037619e7-e17f-4434-9de7-1e535578cd6b","order_by":8,"name":"Maria Reyes Abad Sazatornil","email":"","orcid":"","institution":"Hospital Universitario Miguel Servet","correspondingAuthor":false,"prefix":"","firstName":"Maria","middleName":"Reyes Abad","lastName":"Sazatornil","suffix":""}],"badges":[],"createdAt":"2026-01-22 13:23:23","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8670135/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8670135/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":101943161,"identity":"fd717f90-87cd-459a-9439-8bb8258955db","added_by":"auto","created_at":"2026-02-05 09:40:50","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":3310693,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eIndividual predicted concentrations vs. observed concentrations among pharmacokinetic models of adalimumab.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe solid line represents the line of identity (\u003cem\u003ey = x\u003c/em\u003e), and the dashed lines indicate the 90% prediction interval. Each dot corresponds to an observed concentration. Pearson’s correlation (r) and significance level are shown.\u003c/p\u003e","description":"","filename":"Fig1.IPREDvs.OBSconcentrations.png","url":"https://assets-eu.researchsquare.com/files/rs-8670135/v1/1dc8b3d51a8fca83c630de91.png"},{"id":101785464,"identity":"fb39284b-7ad8-4a25-87ee-644ffe5bdf3c","added_by":"auto","created_at":"2026-02-03 15:36:39","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":282452,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePrediction-corrected Visual Predictive Checks (pcVPC) among pharmacokinetic models of adalimumab.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe red shaded area represents the 90% prediction interval of the simulated median, while the blue shaded area represents the 90% prediction interval of the simulated 5th and 95th percentiles. Red circles connected by blue lines correspond to the observed percentiles (5th, 50th, and 95th) of the measured concentrations. Coverage below each plot indicates the proportion of observed data falling within the simulated 90% prediction interval.\u003c/p\u003e","description":"","filename":"Fig2.pcVPC.png","url":"https://assets-eu.researchsquare.com/files/rs-8670135/v1/0b678329586fa47e8421cbe9.png"},{"id":101944180,"identity":"78965f92-2bd0-425a-848a-765d3798be49","added_by":"auto","created_at":"2026-02-05 09:49:55","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":4955468,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8670135/v1/f32d26ec-02d9-41e1-b00e-3a2bef9718d5.pdf"},{"id":101785466,"identity":"4de8f3c3-71de-41fe-81e0-b5d481ec52d3","added_by":"auto","created_at":"2026-02-03 15:36:39","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":6591004,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementarymaterial.docx","url":"https://assets-eu.researchsquare.com/files/rs-8670135/v1/d02fbe77bb40adecb5e9a3ee.docx"},{"id":101785467,"identity":"6700320e-192f-4f07-adf4-8c6e7231cfcc","added_by":"auto","created_at":"2026-02-03 15:36:39","extension":"tiff","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":12505134,"visible":true,"origin":"","legend":"","description":"","filename":"SlFig4.Analysisoftheresiduals.tiff","url":"https://assets-eu.researchsquare.com/files/rs-8670135/v1/2fb6a5004f7f638048dca1fa.tiff"}],"financialInterests":"No competing interests reported.","formattedTitle":"External validation of population pharmacokinetic models of adalimumab in adult patients with inflammatory bowel disease: Towards model-informed precision dosing","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eInflammatory bowel disease (IBD), which includes Crohn\u0026rsquo;s disease (CD) and ulcerative colitis (UC), is a chronic disease characterized by significant morbidity and impaired quality of life [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. Biologic therapies, in particular adalimumab (ADA), the most used fully human monoclonal antibody directed against tumor necrosis factor alpha (anti-TNFα), have substantially advanced the management of moderate-to-severe CD and UC by achieving and sustaining remission [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eRegarding to ADA, the most common treatment strategy consists of administering in a standardized dosage regimen, which does not take into account the differences between individuals and the variability in drug metabolism and elimination [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. Notably, the pharmacokinetics of ADA in patients with IBD are highly variable so standard dosing may lead to suboptimal plasma levels [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. In turn, low ADA concentrations have been associated with increased risk of treatment failure and development of anti-adalimumab antibodies (AAA) [\u003cspan additionalcitationids=\"CR6 CR7 CR8\" citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. These limitations have led to empirical adjustments in dose and/or frequency of administration in clinical practice to optimize clinical response and to ensure the effectiveness and safety of treatment [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eAccordingly, population pharmacokinetic (popPK) models enable the quantification of pharmacokinetic variability among patients treated with ADA, as they incorporate individual variables such as gender, weight, C-reactive protein (CRP), faecal calprotectin (FCP), and albumin, among others. Model-Informed Precision Dosing (MIPD) based on popPK models is a tool that allows treatment optimization using Bayesian statistical estimates that predict ADA distribution and elimination for each patient and help adjust doses and/or dosing intervals [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. In this way, the MIPD seeks to individualize treatment dosages to promote the rational use of the drug, maximizing its safety and effectiveness [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. Its use through therapeutic drug monitoring (TDM) and analysis of AAA in serum has become a recognized clinical strategy [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. Although evidence remains scarce, many studies point to its usefulness in reducing therapeutic failure and towards early optimization of ADA dosing [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eHowever, given that popPK models are developed using distinct patient cohorts, their predictive performance may vary across different clinical settings. In addition, popPK models have generally been conducted in small cohorts, in different geographical contexts and often excluding patients with UC, making it essential to conduct external validation studies in larger samples that are representative of actual clinical practice. In this context, the aim of this study was to compare the main adult IBD patients popPK models of adalimumab in order to analyze their predictive performance in a representative and significant sample of a tertiary hospital. As an exploratory analyses, we also assessed the associations between HLA-DQA1*05 and ADA concentrations.\u003c/p\u003e"},{"header":"2. Methodology","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Study design and participants\u003c/h2\u003e \u003cp\u003eThis was a retrospective and longitudinal study, performed at a tertiary care hospital, that consecutively recruited all IBD patients treated with ADA between January/2022 to October/2024 for whom at least two serum measurements of ADA concentration are available. The reporting of this study was conducted in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement. Demographic, clinical and pharmacokinetic data were collected from medical records. Patients for whom the necessary information to perform the pharmacokinetic analysis in the different study models was not available were excluded. ADA treatment consisted in an induction phase that was established as 160mg and 80mg at weeks 0 and 2 respectively. Following this phase, as a maintenance phase, patients were treated with 40mg of adalimumab every other week. If clinically indicated and under medical prescription, treatment intensification was allowed.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Laboratory analysis\u003c/h2\u003e \u003cp\u003eEach sample was collected before the corresponding dose of ADA during induction and/or maintenance phase (trough levels). ADA serum concentrations were measured using an enzyme-linked immunosorbent assay (Promonitor\u0026reg;-ADA, Grifols Diagnostic Solutions Inc., California, US). The limits of quantification were: 0.25 and 12.0\u0026micro;g/ml. For samples above the upper limit, additional dilution was required. AAA determination was performed using a drug-sensitive technique, consequently, AAA detection was limited to samples lacking detectable ADA. Therefore, the presence of AAA was only evaluated upon request by the clinician in patients with therapeutic failure and a high suspicion of immunogenicity, defined as unexplained decline in adalimumab serum concentrations.\u003c/p\u003e \u003cp\u003eADA concentrations\u0026thinsp;\u0026ge;\u0026thinsp;5 \u0026micro;g/mL and \u0026ge;\u0026thinsp;6.4 \u0026micro;g/mL at induction phase were considered within the therapeutic range for CD and UC, respectively. Therapeutic levels during the maintenance phase were defined as ADA concentrations\u0026thinsp;\u0026gt;\u0026thinsp;10\u0026ndash;12 \u0026micro;g/mL, associated with endoscopic healing and clinical remission [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eScreening of HLA-DQA1*05 genotype was realized with the polymerase chain reaction method, using the CeliacStrip\u0026reg; test (Operon S.A., Zaragoza, Spain). Other laboratory variables necessary for clinical follow-up were also collected, such as CRP, FCP, serum albumin, hemoglobin and erythrocyte sedimentation rate.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Validation of models\u003c/h2\u003e \u003cp\u003eSix adult popPK models were selected: Berends, FDA, Marquez, Sanchez, Ternant and Vande, respectively [\u003cspan additionalcitationids=\"CR19 CR20 CR21 CR22\" citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. All of them describe ADA pharmacokinetics using a one-compartment structure with first-order absorption in adult patients. ADA popPK models and equations are shown in \u003cem\u003eSupplementary Tables\u0026nbsp;1 and 2\u003c/em\u003e.\u003c/p\u003e \u003cp\u003eThe predictive performance of every model was evaluated by visual analysis of the goodness-of-fit (GOF) plots, and prediction-corrected visual predicted checks (pcVPC). In GOF plots a good predictive performance was defined as a close alinement of the data points along the identity line. The correlation in GOF plots is presented solely as a descriptive measure. Coverage, defined as the proportion of observed ADA serum concentrations falling within the simulated 90% prediction intervals, was computed for every model in line with the pcVPC. A good predictive performance was considered to occur when the empirical percentiles were contained within these intervals, and higher coverage values indicated better agreement between observed and simulated data.\u003c/p\u003e \u003cp\u003eWe calculated the distribution of the Empirical Bayesian estimate (EBEs) of the pharmacokinetic parameters for each of the popPK models. This data was compared with the theoretical distribution of these pharmacokinetic parameters according to every popPK model.\u003c/p\u003e \u003cp\u003eBias and precision of the models were calculated through mean prediction error (MPE), mean absolute prediction error (MAPE) and root mean square error (RMSE) based on the following formulas: in which \u003cem\u003eỸ\u003c/em\u003e represents the population predicted ADA concentration, while \u003cem\u003eY\u003c/em\u003e represents the observed ADA concentration and \u003cem\u003en\u003c/em\u003e is the total of observations analyzed.\u003cdiv id=\"Equa\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equa\" name=\"EquationSource\"\u003e\n$$\\:MPE=\\frac{\\sum\\:(Ỹ-Y)}{n}\\:\\:\\:\\:\\:\\:\\:\\:\\:\\:\\:\\:\\:MAPE=\\left|\\frac{\\sum\\:(Y\\:-Ỹ)}{Ỹ}\\right|x\\frac{100}{n}\\:RMSE=\\sqrt{\\frac{{\\sum\\:(Ỹ-Y)}^{2}}{n}}$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eThis study used a bootstrap resampling approach to compare the statistical behavior of the three metrics employed. Baseline metrics were first calculated using paired observations of actual and predicted values; then 1000 bootstrap samples were generated by random resampling with replacement. Each bootstrap sample was used to recalculate the metrics, thereby obtaining empirical distributions from which means, standard deviations, and 95% confidence intervals were estimated. The statistical comparisons were adjusted for multiple comparisons using Bonferroni correction to avoid spurious statistical interpretations of the results. Further, the structural models were compared using Akaike Information Criterion (AIC), with lower AIC values indicating better model fit. Such a comparison of both the metric distributions and AIC values allowed for a robust assessment of the variability and bias, ensuring a reproducible framework for performance evaluation.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e2.4 Statistical analysis\u003c/h2\u003e \u003cp\u003eThe Shapiro-Wilk test and P-P plots were used to assess the distribution of variables. Comparisons between variables were made using Chi-square statistical tests for qualitative variables; Student\u0026rsquo;s t-test for comparison between means if normally distributed variables; Mann-Whitney test, Wilcoxon test and Kruskal-Wallis test were used for comparisons between groups with non-parametric distribution. The correlation between observed and individual predicted values was measured using a Pearson\u0026rsquo;s correlation analysis. A p-value of \u0026lt;\u0026thinsp;0.05 was considered significant. Analyses were carried out with STATA/MP\u0026reg; v.17 and R-Studio\u0026reg; v.2024.12.0. Simulation and estimation of pharmacokinetic models were performed using MonolixSuite\u0026trade; v.2024.R1.\u003c/p\u003e \u003c/div\u003e"},{"header":"3. Results","content":"\u003cp\u003e\u003cstrong\u003e3.1 Sample characteristics\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA total of 201 patients were included in the study, of whom 177 (88.1%) had Crohn\u0026rsquo;s disease (CD). Females represented 44.8%, the mean basal age was 48.6\u0026plusmn;16.2 years and mean body mass index (BMI) was 25.3\u0026plusmn;5.2 kg/m\u003csup\u003e2\u003c/sup\u003e, with no significant difference between CD and UC participants (\u003cem\u003ep\u003c/em\u003e \u0026gt; 0.126). Years of disease progression were significantly longer in UC patients (14.8; IQR: [4.7-19.5]) than in CD (5.8; IQR [2.4-14.2]; \u003cem\u003ep\u003c/em\u003e = 0.018) at the start of pharmacokinetic monitorization. The majority of patients (72.6%) received treatment with immunomodulators. Although these were mainly prescribed prior to receiving adalimumab, 27.9% of the subjects received them concomitantly with ADA. The most commonly used immunomodulator was azathioprine (66.2%). Methotrexate was only used among CD patients (14.4%). Prior anti-TNF\u0026alpha; therapy had been administered to 11.9% of the cohort, predominantly infliximab. The rest of data are shown in \u003cem\u003eTable 1.\u003c/em\u003e Data according to Montreal classification is shown in \u003cem\u003eSupplementary Table 3. \u0026nbsp;\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eA total of 709 ADA serum samples were analyzed, mostly during maintenance phase (74.8%). The average amount of ADA serum samples per subject was 3 (IQR [2-4]). The overall mean ADA serum concentration was 13.8\u0026plusmn;4.5 mg/L at week 2, 13.4\u0026plusmn;6.2 mg/L at week 6 and 12.3\u0026plusmn;7.1 mg/L during maintenance phase. In both the induction and maintenance phases, no differences in ADA concentrations were observed between EC and UC. However, a significantly higher proportion of patients with UC had ADA concentrations within the therapeutic target range compared to those with CD (80.5% vs. 64.8%; p=0.004). This is probably why, mainly in the CD group the dosage regimen was increased to a weekly basis in 14.4% of all patients, while 6.2% of the subjects had their maintenance dose increased to 80 mg. No differences were observed in ADA concentrations in patients with perianal disease (p=0.12).\u0026nbsp;Overall, AAA were detected in 5.5% of the samples, all of which occurred in CD patients.\u003c/p\u003e\n\u003cp\u003eAn exploratory analysis was conducted to compare groups with and without HLA-DQA1*05. The allele was present in 38% of the evaluated patients, without significant differences between CD and UC groups (\u003cem\u003ep\u003c/em\u003e = 1.000). ADA concentrations within the therapeutic range were achieved by 67% of individuals with HLA-DQA1*05 and by 70.7% of patients without the allele (p=0.215). Neither were there any significant differences in ADA serum concentrations (12.8\u0026plusmn;6.9 mg/L and 12.4\u0026plusmn;5.5 mg/L for patients without and with the allele, respectively). Only 6 of the patients with AAA were tested for the presence of HLA-DQA1*05. No association was observed between HLA-DQA1*05 and AAA formation (1.07% and 1.19% for the absent and present allele, respectively, p=0.631).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.2 Validation of models\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAs shown in \u003cem\u003eFigure 1\u003c/em\u003e, the correlation between predicted and observed adalimumab concentrations varied substantially among the compared popPK models. Of these, the Vande and Berends models displayed the strongest correlations, at r=0.83 and r=0.77 (p\u0026lt;0.001 for both), respectively, with data points in close proximity to the line of identity. In contrast, the Sanchez and Ternant\u0026rsquo;s models exhibited weak to moderate correlations (r=0.45 and r=0.37; p\u0026lt;0.001), representing the poorest correlations among the models evaluated and the largest scatter around the line of identity. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThese findings were further confirmed by the predictive check with pcVPC plots presented in \u003cem\u003eFigure 2\u003c/em\u003e. The model of Berends showed the best performance: more than 90% of the observed data fell within the 90% prediction interval. The models of Vande and FDA also performed well, with more than 80% of observations covered. In contrast, for all the other models, a considerable amount of data fell outside of the prediction intervals, reflecting systematic bias and lower precision. More importantly, all models provided a negative prediction of the central tendency; thus, the simulated medians were consistently lower than the respective observed median values. As shown in \u003cem\u003eTable 2\u003c/em\u003e, ADA clearance was generally overestimated in most models, except for S\u0026aacute;nchez and Ternant\u0026rsquo;s models, whose population and individual estimates were closely aligned.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThese observations are also consistent with the performance metrics outlined in \u003cem\u003eTable 3\u003c/em\u003e and \u003cem\u003eSupplementary Figures 1-3\u003c/em\u003e. The MPE showed significant differences for most models at p \u0026lt; 0.001, except between Sanchez and Ternant\u0026rsquo;s models, which were comparably poor in their predictive abilities (p = 0.112). Conversely, the models of Berends and Vande had respective MPE values of 0.58 and \u0026ndash;0.43 \u0026micro;g/mL, which reflected minimal bias and a strong agreement between the predicted and observed values.\u003c/p\u003e\n\u003cp\u003eInterestingly, in the Sanchez\u0026rsquo;s model, the MAPE was the lowest at 1.37% (95%CI: 1.32-6.71), indicating that it was relatively more accurate in its predictions. However, its MAPE values were more spread out indicating more variability in the individual estimates. The Marquez\u0026rsquo;s model had a relatively low MAPE of 1.86% (95%CI: 0.45-3.28), consistent with its modest MPE, indicating an acceptable model fit. Finally, the model developed by FDA had the highest MAPE 7.0% (2.35-11.65) and was thus the least reliable.\u003c/p\u003e\n\u003cp\u003eIn terms of RMSE comparison between models, the Vande and Berends models were again the best (3.85 \u0026micro;g/mL (95%CI: 3.55-4.14) and 4.68 \u0026micro;g/mL (95%CI: 4.19-5.16) \u0026micro;g/mL, respectively), reflecting the highest predictive precision. The poorest performances regarding the RMSE of estimates included Sanchez and Ternant models 7.51\u0026micro;g/mL (95%CI: 6.78-8.24) and 7.84\u0026micro;g/mL (95%CI: 6.99-8.71), respectively. Moreover, the Berends and Vande models also demonstrated the lowest AIC values, reinforcing their superior model fit.\u003c/p\u003e\n\u003cp\u003eResidual distribution in both the Berends and Vande\u0026rsquo;s models resembled normality, was centered at zero, and showed homogeneous dispersion, which suggests good overall fit and absence of systematic bias. Graphs of residuals versus individual prediction and versus time showed that in both models, random dispersion without structured patterns was maintained throughout, supporting consistency of the model across a wide range of concentrations and follow-up time. By contrast, the models from the FDA, Marquez, Sanchez, and Ternant had greater dispersion, were heteroscedastic, with residuals further away from zero with noticeable clusters in some time periods or prediction levels. This analysis suggests lower predictive capacity in more variable clinical scenarios. Figures are shown in \u003cem\u003eSupplementary Figure 4\u003c/em\u003e. \u0026nbsp;\u003c/p\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eIn the current study, we performed an extensive external validation of six published popPK models of ADA in a cohort of 201 IBD patients. We compared their predictive performance to identify which model best describes ADA pharmacokinetics in this population. Our results consistently confirm the models of Berends and Vande as most reliable in predicting adalimumab pharmacokinetics within this population. Given that they appear superior across multiple metrics for evaluation, their superiority is indicative of the best balance between bias, precision, and overall predictive accuracy.\u003c/p\u003e \u003cp\u003eThe superiority of Berends [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e] and Vande [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e] models was evident in highest closeness to the line of identity in GOF plots and favorable pcVPC plots, with more than 80% of observed values falling within the 90% prediction interval. These findings indicate a good match between model-derived predictions and real-world ADA concentrations. In addition, both models showed low MPE, low RMSE, and homogeneously distributed residuals centered around zero, key indicators of low bias and high precision [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e]. However, the simulated median tended to underestimate the observed median concentrations, which is partly justified by the overestimation of ADA population clearance. Nevertheless, this bias is small and does not compromise the utility of these popPK models for MIPD, as Bayesian updating corrects for population-level discrepancies when patient-specific data are incorporated.\u003c/p\u003e \u003cp\u003eThe suitability of the Berends model is consistent with a study conducted by Marquez-Megias S, et al [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. Although the population studied differs from ours, as it evaluates patients with biomarkers indicative of a greater IBD activity (with lower serum albumin, higher CRP and FCP), the suitability of Berends [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e] model is probably due to the selected model's covariates. As shown in \u003cem\u003eSupplementary Table\u0026nbsp;2\u003c/em\u003e, Berends model incorporates the dosing regimen (every other week vs. every week), which is particularly relevant given that in our population 23.4% of subjects receive every week regimen. This covariate allows the effect of the regimen to be reflected on clearance, improving the model's ability to reproduce the concentrations observed in typical clinical situations of treatment optimization. In addition, Vande's model incorporates the lean body weight (LBW) as a main covariate. We believe that LBW is particularly appropriate for our cohort since it is normally distributed and do not include extreme values, unlike other variables proposed by other models (such as FCP or serum albumin etc.). Consequently, the relationship between LBW and ADA clearance is more consistent and less affected by outliers. This contributes to LBW being a robust covariate and helps to improve the predictive capacity of the model in comparison to other variables.\u003c/p\u003e \u003cp\u003eIn contrast, the models by Sanchez [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e] and Ternant [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e], despite estimating clearance values similar to those in our cohort, demonstrated poor predictive performance. Although in 11.9% of our patients ADA immunogenicity was suspected, tests to detect the presence of AAA were only performed in 11 of them, resulting in a low prevalence of positive AAA in our cohort (5.5%). The large impact of AAA on ADA clearance may have contributed to the robustness of Berends [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e] and Vande [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e] models, but Ternant model [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e], by proposing only this variable as a predictor of ADA clearance, results in most patients in our study having a fixed clearance, leading to no significant differences between theoretical and estimated clearance, but without the ability to reproduce the actual variability between patients. This lack of explained variability is reflected in the poor alignment of individual predictions and insufficient pcVPC. It is likely that if the proportion of AAAs evaluated were higher in our subjects, Ternant model would predict better, as is the case in other studies [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. Contrary, in Sanchez model [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e], this discrepancy is probably due to its lack of parsimony. The structural complexity and high number of covariates may have led to overfitting in the original population, limiting its extrapolation capacity and explaining the poor individual fit and low coverage of the pcVPC in our study.\u003c/p\u003e \u003cp\u003eThe FDA model [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e] also performed reasonably well. However, its lack of covariates explaining clearance, limits its ability to capture interindividual variability in ADA clearance. In contrast, the model developed by Marquez [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e] was based on patients with a presumably more severe disease activity, characterized by lower albumin, higher FCP and CRP levels. These differences in baseline characteristics may be partly responsible for the lower accuracy observed when applying these models to our dataset.\u003c/p\u003e \u003cp\u003eIn our cohort, the presence of the HLA-DQA1*05 allele was not associated with differences in serum ADA concentrations, in the proportion of patients who reached therapeutic levels, or in the formation of AAA. These results differ from previous studies that have associated the allele with an increased immunogenicity against anti-TNFα and subtherapeutic concentrations of anti-TNFα, but no greater therapeutic failure [\u003cspan additionalcitationids=\"CR27\" citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]. Our results could be explained by the small sample size, especially in the AAA group, and by the high proportion of patients with empirically intensified regimens (33.8%).\u003c/p\u003e \u003cp\u003e This study presents several strengths, including the relatively large cohort, a high number of validated samples per subject, and its conduct in real-world clinical practice involving patients with IBD represents an unselected sample of all patients who initiated ADA at our hospital. All popPK models evaluated were developed in patients with IBD, in contrast to most previous studies that have attempted to extrapolate popPK models developed for other autoimmune diseases, such as psoriasis or rheumatoid arthritis, to populations with IBD [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]. In addition, the methodology used is consistent and rigorous, with multiple validation metrics employed. This facilitates the extrapolation of the findings to other contexts, thereby enhancing the external validity of the study. Nonetheless, some limitations must be acknowledged. First, the retrospective design may introduce biases related to sampling time and in data collection. Second, we lacked data on baseline disease activity, so related secondary variables were collected.\u003c/p\u003e \u003cp\u003eWe consider the external validation approach used in the study an appropriate first step toward implementing MIPD in clinical practice, facilitating individualized ADA dosing in patients with IBD. This is particularly relevant for CD patients, who in our study exhibited a higher frequency of subtherapeutic ADA levels compared to those with UC. Future research should focus on conducting prospective studies, exploring how other covariates, such as HLA-DQA1*05, influence ADA clearance, and developing popPK models with sufficient external validity to allow their extrapolation to different patients and clinical contexts.\u003c/p\u003e \u003cp\u003eIn conclusion, the superior predictive performance of the Berends and Vande models underscores their pivotal role in modernizing adalimumab therapy. Our results provide a solid foundation for implementing MIPD, moving beyond empirical adjustments. Therefore, pending further evidence from diverse populations, it is crucial to prioritize the adoption of these models to maximize therapeutic success and minimize bias in clinical decision-making.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eFunding:\u003c/strong\u003e No funding was received to assist with the preparation of this manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests:\u003c/strong\u003e The authors declare no conflicts of interest.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Availability Statement:\u003c/strong\u003e The datasets generated during the current study are available from the corresponding author on reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval:\u0026nbsp;\u003c/strong\u003eThe study was approved by the Clinical Research Ethics Committee of Autonomous Community of Aragon (CEICA) with the following registration number: EOM24/034. The procedures established in this research are designed following the principles of good clinical practice and the declaration of Helsinki.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eInformed consent:\u003c/strong\u003e All participants provided written informed consent before enrollment in the study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions:\u0026nbsp;\u003c/strong\u003eAll authors contributed to the study conception and design. Data collection and analysis were performed by Irene Aguilo-Lafarga, Tonet Ser\u0026eacute;s-Noriega, Vicente Gimeno Ballester and Pilar Antonia Corsino Roche. The first draft of the manuscript was written by Irene Aguilo-Lafarga. \u0026nbsp;Tonet Ser\u0026eacute;s-Noriega, Vicente Gimeno Ballester, Pilar Antonia Corsino Roche, Raquel Vicente Lidon, Eva Maria Sierra Moros, Sonia Gallego, Santiago Garcia Lopez and Maria Reyes Abad Sazatornil reviewed the relevant literature and provided a critical review of the manuscript.\u0026nbsp;\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eMoran GW, Gordon M, Sinopolou V, et al (2025) British Society of Gastroenterology guidelines on inflammatory bowel disease in adults: 2025. Gut 74:s1\u0026ndash;s101. https://doi.org/10.1136/gutjnl-2024-334395\u003c/li\u003e\n\u003cli\u003eGordon H, Minozzi S, Kopylov U, et al (2024) ECCO Guidelines on Therapeutics in Crohn\u0026rsquo;s Disease: Medical Treatment. J Crohns Colitis 18:1531\u0026ndash;1555. https://doi.org/10.1093/ecco-jcc/jjae091\u003c/li\u003e\n\u003cli\u003eRaine T, Bonovas S, Burisch J, et al (2022) ECCO Guidelines on Therapeutics in Ulcerative Colitis: Medical Treatment. J Crohns Colitis 16:2\u0026ndash;17. https://doi.org/10.1093/ecco-jcc/jjab178\u003c/li\u003e\n\u003cli\u003eVande Casteele N, Gils A. (2015) Pharmacokinetics of anti-TNF monoclonal antibodies in inflammatory bowel disease: Adding value to current practice. J Clin Pharmacol. 55:S39\u0026ndash;50. https://doi.org/10.1002/jcph.374\u003c/li\u003e\n\u003cli\u003eRoblin X, Marotte H, Rinaudo M, et al (2014) Association between pharmacokinetics of adalimumab and mucosal healing in patients with inflammatory bowel diseases. Clin Gastroenterol Hepatol 12:80-84.e2. https://doi.org/10.1016/j.cgh.2013.07.010\u003c/li\u003e\n\u003cli\u003eRueda Garc\u0026iacute;a JL, Su\u0026aacute;rez-Ferrer C, Amiama Roig C, et al (2025) Association of early therapeutic drug monitoring of adalimumab with biologic remission and drug survival in Crohn\u0026rsquo;s Disease. Therap Adv Gastroenterol 18:17562848251324226. https://doi.org/10.1177/17562848251324226\u003c/li\u003e\n\u003cli\u003eYanai H, Lichtenstein L, Assa A, et al (2015) Levels of drug and antidrug antibodies are associated with outcome of interventions after loss of response to infliximab or adalimumab. Clin Gastroenterol Hepatol 13:522-530.e2. https://doi.org/10.1016/j.cgh.2014.07.029\u003c/li\u003e\n\u003cli\u003eUngar B, Levy I, Yavne Y, et al. (2016) Optimizing Anti-TNF-\u0026alpha; Therapy: Serum Levels of Infliximab and Adalimumab Are Associated With Mucosal Healing in Patients With Inflammatory Bowel Diseases. Clin Gastroenterol Hepatol. 14:550-557.e2. https://doi.org/10.1016/j.cgh.2015.10.025\u003c/li\u003e\n\u003cli\u003eKennedy NA, Heap GA, Green HD, et al (2019) Predictors of anti-TNF treatment failure in anti-TNF-naive patients with active luminal Crohn\u0026rsquo;s disease: a prospective, multicentre, cohort study. Lancet Gastroenterol Hepatol 4:341\u0026ndash;353. https://doi.org/10.1016/S2468-1253(19)30012-3\u003c/li\u003e\n\u003cli\u003eMa C, Huang V, Fedorak DK, et al (2014) Adalimumab dose escalation is effective for managing secondary loss of response in Crohn\u0026rsquo;s disease. Aliment Pharmacol Ther 40:1044\u0026ndash;1055. https://doi.org/10.1111/apt.12940\u003c/li\u003e\n\u003cli\u003eSandborn WJ, Colombel J-F, Schreiber S, et al (2011) Dosage adjustment during long-term adalimumab treatment for Crohn\u0026rsquo;s disease: clinical efficacy and pharmacoeconomics. Inflamm Bowel Dis 17:141\u0026ndash;151. https://doi.org/10.1002/ibd.21328\u003c/li\u003e\n\u003cli\u003eMinichmayr IK, Dreesen E, Centanni M, et al (2024) Model-informed precision dosing: State of the art and future perspectives. Adv Drug Deliv Rev 215:115421. https://doi.org/10.1016/j.addr.2024.115421\u003c/li\u003e\n\u003cli\u003ePapamichael K, Stocco G, Ruiz Del Agua A (2023) Challenges in Therapeutic Drug Monitoring: Optimizing Biological Treatments in Patients With Inflammatory Bowel Disease and Other Immune-Mediated Inflammatory Diseases. Ther Drug Monit 45:579\u0026ndash;590. https://doi.org/10.1097/FTD.0000000000001095\u003c/li\u003e\n\u003cli\u003eHanauer SB, Torres EA, Aragon-Han P, et al (2025) The Clinical Utility of Precision-Guided Dosing for Adalimumab Therapy Optimization in Inflammatory Bowel Disease: A Clinical Experience Program. Pharmaceutics 17:428. https://doi.org/10.3390/pharmaceutics17040428\u003c/li\u003e\n\u003cli\u003ePapamichael K, Cheifetz AS, Melmed GY, et al (2019) Appropriate Therapeutic Drug Monitoring of Biologic Agents for Patients With Inflammatory Bowel Diseases. Clin Gastroenterol Hepatol 17:1655-1668.e3. https://doi.org/10.1016/j.cgh.2019.03.037\u003c/li\u003e\n\u003cli\u003eRodr\u0026iacute;guez-Moranta F, Arg\u0026uuml;elles-Arias F, Hinojosa Del Val J, et al (2024) Therapeutic drug monitoring in inflammatory bowel diseases. Position statement of the Spanish Working Group on Crohn\u0026rsquo;s Disease and Ulcerative Colitis. Gastroenterol Hepatol 47:522\u0026ndash;552. https://doi.org/10.1016/j.gastrohep.2024.01.007\u003c/li\u003e\n\u003cli\u003ePapamichael K, Juncadella A, Wong D, et al (2019) Proactive Therapeutic Drug Monitoring of Adalimumab Is Associated With Better Long-term Outcomes Compared With Standard of Care in Patients With Inflammatory Bowel Disease. J Crohns Colitis 13:976\u0026ndash;981. https://doi.org/10.1093/ecco-jcc/jjz018\u003c/li\u003e\n\u003cli\u003eBerends SE, Mould DR, Strik AS, et al. (2018) Explaining Interpatient Variability in Adalimumab Pharmacokinetics in Patients With Crohn\u0026rsquo;s Disease. Ther Drug Monit.40:202\u0026ndash;11. https://doi.org/10.1097/ftd.0000000000000494\u003c/li\u003e\n\u003cli\u003eFood and Drug Administration (FDA) (2008). Humira: Clinical Pharmacology and Biopharmaceutics Review(s). FDA. Available online: https://www.accessdata.fda.gov/drugsatfda_docs/bla/2007/125057_S0089.pdf. Accessed 28 Nov 2025.\u003c/li\u003e\n\u003cli\u003eMarquez-Megias S, Nalda-Molina R, M\u0026aacute;s-Serrano P, Ramon-Lopez A. (2023) Population Pharmacokinetic Model of Adalimumab Based on Prior Information Using Real World Data. Biomedicines. 11:2822. https://doi.org/10.3390/biomedicines11102822\u003c/li\u003e\n\u003cli\u003eS\u0026aacute;nchez-Hern\u0026aacute;ndez JG, P\u0026eacute;rez-Blanco JS, Rebollo N, et al (2020) Biomarkers of disease activity and other factors as predictors of adalimumab pharmacokinetics in inflammatory bowel disease. Eur J Pharm Sci 150:105369. https://doi.org/10.1016/j.ejps.2020.105369\u003c/li\u003e\n\u003cli\u003eTernant D, Karmiris K, Vermeire S, et al. (2015) Pharmacokinetics of adalimumab in Crohn\u0026rsquo;s disease. Eur J Clin Pharmacol. 71:1155\u0026ndash;7. https://doi.org/10.1007/s00228-015-1892-1\u003c/li\u003e\n\u003cli\u003eVande Casteele N, Baert F, Bian S, et al (2019) Subcutaneous Absorption Contributes to Observed Interindividual Variability in Adalimumab Serum Concentrations in Crohn\u0026rsquo;s Disease: A Prospective Multicentre Study. J Crohns Colitis 13:1248\u0026ndash;1256. https://doi.org/10.1093/ecco-jcc/jjz050\u003c/li\u003e\n\u003cli\u003eSherwin CMT, Kiang TKL, Spigarelli MG, Ensom MHH (2012) Fundamentals of population pharmacokinetic modelling: validation methods. Clin Pharmacokinet 51:573\u0026ndash;590. https://doi.org/10.1007/BF03261932\u003c/li\u003e\n\u003cli\u003eMarquez-Megias S, Ramon-Lopez A, M\u0026aacute;s-Serrano P, et al (2021) Evaluation of the Predictive Performance of Population Pharmacokinetic Models of Adalimumab in Patients with Inflammatory Bowel Disease. Pharmaceutics 13:1244. https://doi.org/10.3390/pharmaceutics13081244\u003c/li\u003e\n\u003cli\u003eNavajas Hern\u0026aacute;ndez P, Mouhtar El Halabi S, Gonz\u0026aacute;lez Parra AC, et al (2024) Carriage of the HLA-DQA1⋆05 haplotype is associated with a higher risk of infratherapeutic drug concentration and higher immunogenicity in patients undergoing treatment with anti-TNF for inflammatory bowel disease. Therap Adv Gastroenterol 17:17562848241278145. https://doi.org/10.1177/17562848241278145\u003c/li\u003e\n\u003cli\u003ePascual-Oliver A, Casas-Deza D, Cuar\u0026aacute;n C, et al (2024) HLA-DQA1*05 Was Not Associated With Primary Nonresponse or Loss of Response to First Anti-TNF in Real-World Inflammatory Bowel Disease Patients. Inflamm Bowel Dis 30:922\u0026ndash;929. https://doi.org/10.1093/ibd/izad130\u003c/li\u003e\n\u003cli\u003eReppell M, Zheng X, Dreher I, et al (2025) HLA-DQA1*05 Associates With Anti-Tumor Necrosis Factor Immunogenicity and Low Adalimumab Trough Concentrations in Inflammatory Bowel Disease Patients From the SERENE Ulcerative Colitis and Crohn\u0026rsquo;s Disease Studies. J Crohns Colitis 19:jjae129. https://doi.org/10.1093/ecco-jcc/jjae129\u003c/li\u003e\n\u003cli\u003eMart\u0026iacute;nez-Romero GJ, Alvari\u0026ntilde;o A, Hinojosa E, et al (2019) Validation of a population pharmacokinetic model of adalimumab in a cohort of patients with inflammatory bowel disease. Rev Esp Enferm Dig 111:431\u0026ndash;436. https://doi.org/10.17235/reed.2019.5600/2018\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003e\u003cstrong\u003eTable 1. Baseline clinical characteristics of the sample according to the type of inflammatory bowel disease\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"577\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 113px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eGlobal (n=201)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCrohn\u0026rsquo;s disease (n=177)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 94px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eUlcerative colitis (n=24)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 48px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003ep\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e-value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 217px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSex (female)\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 113px;\"\u003e\n \u003cp\u003e88 (44.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e74 (41.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 94px;\"\u003e\n \u003cp\u003e14 (58.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 48px;\"\u003e\n \u003cp\u003e0.126\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 217px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eBasal age (years)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 113px;\"\u003e\n \u003cp\u003e48.6 (\u0026plusmn;16.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e47.9 (\u0026plusmn;16.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 94px;\"\u003e\n \u003cp\u003e53.8 (\u0026plusmn;17.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 48px;\"\u003e\n \u003cp\u003e0.951\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 217px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eIBD duration (years)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 113px;\"\u003e\n \u003cp\u003e6.1 (2.4-15.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e5.8 (2.4-14.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 94px;\"\u003e\n \u003cp\u003e14.8 (4.7-19.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 48px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.018\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 217px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eWeight (kg)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 113px;\"\u003e\n \u003cp\u003e72.2 (\u0026plusmn;17.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e71.9 (\u0026plusmn;16.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 94px;\"\u003e\n \u003cp\u003e70.9 (\u0026plusmn;13.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 48px;\"\u003e\n \u003cp\u003e0.383\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 217px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eHeight (cm)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 113px;\"\u003e\n \u003cp\u003e168.5 (\u0026plusmn;9.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e168.6 (\u0026plusmn;9.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 94px;\"\u003e\n \u003cp\u003e167.8 (\u0026plusmn;10.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 48px;\"\u003e\n \u003cp\u003e0.354\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 217px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eBMI (kg/m\u003csup\u003e2\u003c/sup\u003e)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 113px;\"\u003e\n \u003cp\u003e25.3 (\u0026plusmn;5.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e25.2 (\u0026plusmn;5.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 94px;\"\u003e\n \u003cp\u003e25.0 (\u0026plusmn;3.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 48px;\"\u003e\n \u003cp\u003e0.447\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 217px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eLBW (kg)\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 113px;\"\u003e\n \u003cp\u003e49.7 (\u0026plusmn;0.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e49.8(\u0026plusmn;0.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 94px;\"\u003e\n \u003cp\u003e49.6(\u0026plusmn;1.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 48px;\"\u003e\n \u003cp\u003e0.445\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 217px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eBSA (m\u003csup\u003e2\u003c/sup\u003e)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 113px;\"\u003e\n \u003cp\u003e1.82 (\u0026plusmn;0.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e1.83 (\u0026plusmn;0.24)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 94px;\"\u003e\n \u003cp\u003e1.81 (\u0026plusmn;0.22)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 48px;\"\u003e\n \u003cp\u003e0.397\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 217px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAlbumin (g/dL)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 113px;\"\u003e\n \u003cp\u003e4.2 (4.0-4.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e4.2 (4.0-4.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 94px;\"\u003e\n \u003cp\u003e4.2 (4.0-4.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 48px;\"\u003e\n \u003cp\u003e0.120\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 217px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCRP (mg/dL)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 113px;\"\u003e\n \u003cp\u003e0.35 (0.11-0.84)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e0.36 (0.11-0.84)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 94px;\"\u003e\n \u003cp\u003e0.21 (0.10-0.76)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 48px;\"\u003e\n \u003cp\u003e0.584\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 217px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eFCP (mg/kg)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 113px;\"\u003e\n \u003cp\u003e107.8 (50.2-261.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e104.0 (50.2-244.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 94px;\"\u003e\n \u003cp\u003e244.8 (49.1-674.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 48px;\"\u003e\n \u003cp\u003e0.060\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 217px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eHemoglobin (g/dL)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 113px;\"\u003e\n \u003cp\u003e13.9 (\u0026plusmn;1.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e13.9 (\u0026plusmn;1.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 94px;\"\u003e\n \u003cp\u003e13.8 (\u0026plusmn;1.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 48px;\"\u003e\n \u003cp\u003e0.390\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 217px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eESR (mm/h)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 113px;\"\u003e\n \u003cp\u003e12 (5-22)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e12 (5-21)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 94px;\"\u003e\n \u003cp\u003e15(4-32)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 48px;\"\u003e\n \u003cp\u003e0.664\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 217px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eExtraintestinal manifestations\u003c/strong\u003e\u003csup\u003e\u0026int;\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 113px;\"\u003e\n \u003cp\u003e40 (19.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e36 (20.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 94px;\"\u003e\n \u003cp\u003e4 (16.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 48px;\"\u003e\n \u003cp\u003e0.791\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"3\" style=\"width: 113px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eConcomitant immunomodulator treatment\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e6-Mercaptopurine\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 113px;\"\u003e\n \u003cp\u003e13 (6.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e12 (7.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 94px;\"\u003e\n \u003cp\u003e1 (4.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 48px;\"\u003e\n \u003cp\u003e1.000\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003eAzathioprine\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 113px;\"\u003e\n \u003cp\u003e133 (66.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e119 (67.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 94px;\"\u003e\n \u003cp\u003e14 (58.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 48px;\"\u003e\n \u003cp\u003e0.387\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003eMethotrexate\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 113px;\"\u003e\n \u003cp\u003e29 (14.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e29 (16.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 94px;\"\u003e\n \u003cp\u003e0 (0.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 48px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.029\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 217px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePrevious anti-TNF\u0026alpha; treatment\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 113px;\"\u003e\n \u003cp\u003e24 (11.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e21 (11.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 94px;\"\u003e\n \u003cp\u003e3 (12.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 48px;\"\u003e\n \u003cp\u003e1.000\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 217px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePositive HLA-DQA1*05\u003c/strong\u003e\u003csup\u003eǂ\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 113px;\"\u003e\n \u003cp\u003e52 (38.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e47 (37.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 94px;\"\u003e\n \u003cp\u003e5 (38.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 48px;\"\u003e\n \u003cp\u003e1.000\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 217px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePrevious IBD surgery\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 113px;\"\u003e\n \u003cp\u003e47 (23.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e46 (26.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 94px;\"\u003e\n \u003cp\u003e1 (4.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 48px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.017\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 217px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eActive tobacco use\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 113px;\"\u003e\n \u003cp\u003e46 (22.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e44 (24.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 94px;\"\u003e\n \u003cp\u003e2 (8.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 48px;\"\u003e\n \u003cp\u003e0.075\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 217px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eADA serum samples\u003c/strong\u003e\u003csup\u003eꝬ\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 113px;\"\u003e\n \u003cp\u003e709 (100)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e622 (87.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 94px;\"\u003e\n \u003cp\u003e87 (12.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 48px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 217px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eInduction\u003c/strong\u003e\u003csup\u003eꝬ\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 113px;\"\u003e\n \u003cp\u003e179 (25.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e157 (25.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 94px;\"\u003e\n \u003cp\u003e22 (25.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 48px;\"\u003e\n \u003cp\u003e0.980\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 217px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eMaintenance\u003c/strong\u003e\u003csup\u003eꝬ\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 113px;\"\u003e\n \u003cp\u003e530 (74.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e465 (74.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 94px;\"\u003e\n \u003cp\u003e65 (74.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 48px;\"\u003e\n \u003cp\u003e0.980\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 217px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eADA serum concentrations (mg/L)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 113px;\"\u003e\n \u003cp\u003e12.6 (\u0026plusmn;6.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e11.6 (7.6-16.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 94px;\"\u003e\n \u003cp\u003e12.8 (\u0026plusmn;4.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 48px;\"\u003e\n \u003cp\u003e0.319\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 217px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eADA serum concentrations in target range\u003c/strong\u003e\u003csup\u003eꝬ\u003c/sup\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 113px;\"\u003e\n \u003cp\u003e473 (67.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e403 (64.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 94px;\"\u003e\n \u003cp\u003e70 (80.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 48px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.004\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 217px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSuspected immunogenicity\u003csup\u003e֏\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 113px;\"\u003e\n \u003cp\u003e24 (11.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e23 (11.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 94px;\"\u003e\n \u003cp\u003e1 (4.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 48px;\"\u003e\n \u003cp\u003e0.343\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 217px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAAA positive\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 113px;\"\u003e\n \u003cp\u003e11 (5.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 104px;\"\u003e\n \u003cp\u003e11 (6.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 94px;\"\u003e\n \u003cp\u003e0 (0.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 48px;\"\u003e\n \u003cp\u003e0.377\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eData are shown as n (percentage\u0026nbsp;within each column), mean (\u0026plusmn;SD) or median (Q1-Q3).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003ep\u003c/em\u003e values for group comparisons are reported\u003c/p\u003e\n\u003cp\u003eADA: Adalimumab; BMI: Body mass index; BSA: Body surface area; CRP: C-reactive protein; ESR: erythrocyte sedimentation rate; FCP: faecal calprotectin; IBD: inflammatory bowel disease; LBW: lean body weight; TNF\u0026alpha;: tumor necrosis factor \u0026alpha;.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003csup\u003e\u0026int;\u003c/sup\u003e Extraintestinal manifestations included: axial spondyloarthropathy, peripheral arthropathy, erythema nodosum, pyoderma gangrenosum, sweet\u0026apos;s syndrome, oral ulcers, episcleritis, uveitis and primary sclerosing cholangitis.\u003c/p\u003e\n\u003cp\u003e\u003csup\u003eǂ\u003c/sup\u003eAvailable in n = 137; 124 patients with Crohn\u0026apos;s disease and 13 with ulcerative colitis.\u003c/p\u003e\n\u003cp\u003e\u003csup\u003eꝬ\u003c/sup\u003ePercentage within total adalimumab serum samples.\u003csup\u003e\u0026nbsp;\u003c/sup\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003csup\u003e֏\u003c/sup\u003e\u003c/strong\u003eUnexplained decline in adalimumab serum concentrations. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 2. Comparison between theorical and Estimated Bayesian Estimates (EBEs) clearance among models\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"607\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eBerends\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eFDA\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eMarquez\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSanchez\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTernant\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eVande\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTheorical CL/F (L/day)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp\u003e0.32 (0.32-0.32)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp\u003e0.30 (0.30-0.30)\u003csup\u003e\u0026nbsp;*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp\u003e1.22 (0.98-1.66)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp\u003e0.41 (0.35-0.48)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp\u003e0.42 (0.42-0.42)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp\u003e0.36 (0.29-0.43)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eEBEs CL/F (L/day)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp\u003e0.25 (0.19-0.35)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp\u003e0.26 (0.21-0.33)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp\u003e0.36 (0.26-0.52)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp\u003e0.41 (0.35-0.48)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp\u003e0.42 (0.42-0.42)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp\u003e0.26 (0.19-2.51)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ep-value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026lt;0.001\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026lt;0.001\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026lt;0.001\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp\u003e\u0026gt;0.05\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp\u003e\u0026gt;0.05\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 87px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026lt;0.001\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eData are shown as mean (95%CI).\u003c/p\u003e\n\u003cp\u003eCL/F: clearance; EBEs: Estimated Bayesian Estimates\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003csup\u003e*\u003c/sup\u003eFixed CL/F\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 3. Bias, precision (\u0026micro;g/mL) and Akaike Information Criteria of analyzed population PK models\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"567\" class=\"fr-table-selection-hover\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 132px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eMPE\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e(\u0026micro;g/mL)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 130px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eMAPE\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e(%)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 115px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eRMSE\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e(\u0026micro;g/mL)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 115px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAIC\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eBerends\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 132px;\"\u003e\n \u003cp\u003e0.58 (0.24 to 0.93)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 130px;\"\u003e\n \u003cp\u003e2.77 (0.89 to 4.67)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 115px;\"\u003e\n \u003cp\u003e4.68 (4.19 to 5.16)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 115px;\"\u003e\n \u003cp\u003e4470\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eFDA\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 132px;\"\u003e\n \u003cp\u003e-1.18 (-1.53 to -0.84)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 130px;\"\u003e\n \u003cp\u003e7.00 (2.35 to 11.65)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 115px;\"\u003e\n \u003cp\u003e4.78 (4.41 to 5.16)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 115px;\"\u003e\n \u003cp\u003e4729\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eMarquez\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 132px;\"\u003e\n \u003cp\u003e-2.88 (-3.25 to -2.50)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 130px;\"\u003e\n \u003cp\u003e1.86 (0.45 to 3.28)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 115px;\"\u003e\n \u003cp\u003e5.80 (5.27 to 6.33)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 115px;\"\u003e\n \u003cp\u003e5547\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSanchez\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 132px;\"\u003e\n \u003cp\u003e-4.56 (-5.01 to -4.12)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 130px;\"\u003e\n \u003cp\u003e1.37 (1.32 to 6.71)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 115px;\"\u003e\n \u003cp\u003e7.51 (6.78 to 8.24)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 115px;\"\u003e\n \u003cp\u003e13121\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTernant\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 132px;\"\u003e\n \u003cp\u003e-4.76 (-5.22 to -4.30)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 130px;\"\u003e\n \u003cp\u003e4.37 (1.22 to 7.52)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 115px;\"\u003e\n \u003cp\u003e7.84 (6.99 to 8.71)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 115px;\"\u003e\n \u003cp\u003e11923\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eVande\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 132px;\"\u003e\n \u003cp\u003e-0.43 (-0.71 to -0.14)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 130px;\"\u003e\n \u003cp\u003e2.64 (0.65 to 4.63)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 115px;\"\u003e\n \u003cp\u003e3.85 (3.55 to 4.14)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 115px;\"\u003e\n \u003cp\u003e4700\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eData are shown as mean (95%CI). AIC were truncated to whole numbers.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAIC: Akaike Information Criteria; CI: confidence interval; MAPE: mean absolute percentage error; MPE: mean prediction error; RMSE: root mean square error.\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"journal-of-pharmacokinetics-and-pharmacodynamics","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"jopa","sideBox":"Learn more about [Journal of Pharmacokinetics and Pharmacodynamics](http://link.springer.com/journal/10928)","snPcode":"10928","submissionUrl":"https://submission.nature.com/new-submission/10928/3","title":"Journal of Pharmacokinetics and Pharmacodynamics","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Inflammatory Bowel Diseases, Adalimumab, Pharmacokinetics, Precision Medicine, Population Pharmacokinetics","lastPublishedDoi":"10.21203/rs.3.rs-8670135/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8670135/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eInflammatory bowel disease (IBD) has variability in the pharmacokinetics of adalimumab, which may predispose patients to subtherapeutic concentrations, therapeutic failure, and/or drug immunogenicity. Population pharmacokinetic (popPK) models estimate individual pharmacokinetic parameters and allow to personalize therapeutic regimens through Model-Informed Precision Dosing (MIPD).\u003c/p\u003e \u003cp\u003eThis retrospective, longitudinal study evaluated the predictive performance of six population pharmacokinetic models of adalimumab in adult patients with IBD treated at a tertiary hospital. The adequacy and prediction of the models were externally validated using visual analysis of the goodness-of-fit (GOF) plots, prediction-corrected visual predicted checks (pcVPC), analysis of residuals, and statistical metrics such as the Akaike Information Criteria. Bias and prediction were also calculated using mean prediction error, mean absolute percentage error and root mean square error. Bootstrap resampling was applied for statistical comparisons. Pharmacokinetic parameters were estimated using Bayesian methods and compared to theoretical values.\u003c/p\u003e \u003cp\u003eA total of 201 subjects were included, 88% with Crohn\u0026rsquo;s disease, mean age 48.6 (\u0026plusmn;\u0026thinsp;16.2) years, 44.8% women. Overall, Berends and Vande models demonstrated the best predictive performance in the majority of comparative analyses: higher coverage in pcVPC and correlation between observed and predicted values, lower bias and precision values, the lowest AIC, and a homogeneous distribution of residuals. However, both models overestimated adalimumab population clearance.\u003c/p\u003e \u003cp\u003eThese findings support the application of the Berends and Vande models in clinical MIPD strategies. Therefore, pending broader evidence in different populations, driving MIPD adoption is critical to optimize adalimumab regimens and maximize sustained clinical outcomes.\u003c/p\u003e","manuscriptTitle":"External validation of population pharmacokinetic models of adalimumab in adult patients with inflammatory bowel disease: Towards model-informed precision dosing","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-02-03 15:36:34","doi":"10.21203/rs.3.rs-8670135/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-03-10T08:57:49+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-02-23T08:26:18+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"262277904050937124592354382503545463242","date":"2026-02-11T12:51:24+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"134011177548598885154963077549241538571","date":"2026-02-09T11:56:45+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-02-03T12:41:07+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"91531968552713040025454026522614927843","date":"2026-01-29T07:56:45+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-01-29T07:37:50+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-01-23T12:11:33+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-01-23T12:07:12+00:00","index":"","fulltext":""},{"type":"submitted","content":"Journal of Pharmacokinetics and Pharmacodynamics","date":"2026-01-22T12:44:41+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"journal-of-pharmacokinetics-and-pharmacodynamics","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"jopa","sideBox":"Learn more about [Journal of Pharmacokinetics and Pharmacodynamics](http://link.springer.com/journal/10928)","snPcode":"10928","submissionUrl":"https://submission.nature.com/new-submission/10928/3","title":"Journal of Pharmacokinetics and Pharmacodynamics","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"465f0583-e18c-4d7c-87c9-6cbcae85d0e5","owner":[],"postedDate":"February 3rd, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-04-26T06:08:14+00:00","versionOfRecord":[],"versionCreatedAt":"2026-02-03 15:36:34","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8670135","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8670135","identity":"rs-8670135","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}