Md Almomin 1,7, Zrinka Duvnjak 1,2,7, Casper Steenholdt 3,4, Christina Duval 5, Cæcilie Skejø 5, Thea Vestergaard 5, Simon M D Baunwall 5, Mette Julsgaard 5,6,8, Charlotte Kloft 1,2,8
1 Freie Universität Berlin, Institute of Pharmacy, Department of Clinical Pharmacy and Biochemistry (Berlin, Germany), 2 Graduate Research Training program PharMetrX (, Germany), 3 Odense University Hospital, Department of Medical Gastroenterology (, Denmark ), 4 University of Southern Denmark, Department of Clinical Research- Research Unit of Medical Gastroenterology (, Denmark), 5 Department of Hepatology and Gastroenterology, Aarhus University Hospital (, Denmark ), 6 Institute of Clinical Medicine, Health, Aarhus University (, Denmark), 7 Shared first authorship (, ), 8 Shared senior authorship (, )
Introduction:
Pregnancy involves major physiological changes that can alter dynamics of biomarkers commonly used in pre-pregnancy[1–3]. Among these, C-reactive protein (CRP), a sensitive marker of systemic inflammation, and albumin, which can reflect disease severity and systemic inflammation, are commonly used in inflammatory bowel diseases (IBD)[4,5]. In pregnant patients with IBD, interpretation of CRP and albumin is, however, complicated by overlapping effects of disease and pregnancy[6].
Despite their importance, longitudinal trajectories of CRP and albumin during pregnancy and postpartum remain poorly quantified, particularly when comparing healthy and IBD pregnancies. This study aims to characterise the biomarker trajectories using nonlinear mixed-effects modelling (NLME), with a focus on quantifying the differences between healthy and IBD pregnancy.
Methods:
Data from three studies were pooled for analysis, including one healthy pregnancy cohort (41 individuals) and two IBD pregnancy cohorts receiving ustekinumab (42 patients) or vedolizumab (39 patients), respectively (50 Crohn’s disease (CD) and 31 ulcerative colitis (UC) patients)[7–9]. In total, 545 albumin and 537 CRP observations were analysed, including measurements obtained pre-pregnancy (7.5%: IBD samples only), during pregnancy (70.7%), and post-pregnancy (21.8%: IBD samples only). For IBD samples, the majority were taken while patients were in remission (78.8%). Biomarker trajectories were modelled as a function of fertilisation age (FA), representing the true onset of pregnancy, and time since delivery (DT) to characterise postpartum changes. A piecewise structural model was developed to characterise biomarker patterns from pregnancy through six months postpartum, with FA at delivery (DFA) defining the transition between phases. Different empirical functions were considered for each part: polynomials (up to fourth order), logistic, log-logistic, ordinary and sigmoidal Emax, Gompertz, and Weibull functions. Global physician assessment was tested as time-varying covariate on all model parameters, in addition to time-invariant covariates: group (healthy vs IBD), diagnosis (healthy vs. CD vs. UC), age, and baseline body weight (BWT). Different interindividual variability (IIV) and residual unexplained variability (RUV) models were considered. Model selection was based on the objective function value, Akaike information criterion, goodness-of-fit diagnostics, parameter plausibility and their precision. Left-censored CRP data were handled using the M3 method. Analyses were performed using FOCE LAPLACE method in NONMEM (v7.4.3), PsN (v4.8.1), and Pirana (v2.9.6).
Results:
A piecewise quadratic–exponential model best described albumin trajectories during pregnancy and postpartum, demonstrating a progressive decline across gestation: For IBD, typically from 40.7 g/L at pre-pregnancy to 29.3 g/L at delivery, followed by rapid postpartum recovery to pre-pregnancy levels, with half-maximal recovery occurring at 1.37 weeks postpartum. Healthy pregnant individuals had higher baseline albumin concentrations (44.1 vs. 40.7 g/L) and a 20.2% higher linear gestational coefficient. BWT also showed a statistically significant effect on baseline albumin parameter, with higher BWT associated with lower baseline concentrations (~-1.2%/+10 kg). IIV was estimated on the baseline albumin parameter (standard deviation, SD, 2.93 g/L) and on the linear gestational coefficient (SD 0.0569 g/L/week). Additive RUV was used. All parameters were precisely estimated (relative standard error (RSE) for fixed and random effects ≤32% and <50%, respectively). CRP trajectories during pregnancy and postpartum were adequately described using a piecewise quadratic–linear model, demonstrating first a nonlinear increase followed by a subtle late-gestation decline, that in postpartum period linearly decreased to pre-pregnancy value. CRP levels were consistently highest in UC, followed by CD, and lowest in healthy pregnancies (maximum 6.39 mg/L [28.4 weeks] vs. 3.80 mg/L [15.3 weeks]; and 5.98 mg/L vs. 2.90 mg/L at 38 weeks, for UC and healthy individuals, respectively). BWT affected baseline CRP parameter, with the higher BWT associated with higher baseline CRP concentrations. IIV was estimated on the baseline CRP parameter (120% coefficient of variation). Proportional RUV was used. All parameters were estimated with a good precision (RSE<30%). Conclusion: This analysis identified IBD-driven effects in pregnancy in addition to healthy pregnancy-related physiological adaptation. In IBD pregnancy, albumin values at baseline (pre-pregnancy) were typically lower than in healthy pregnancy and showed a less pronounced decline during gestation, whereas mid-gestational rise in CRP was the most pronounced for UC, followed by CD and healthy pregnancy. These results may support the integration of biomarker dynamics into pharmacokinetic and disease-modelling frameworks for pregnancy/postpartum in IBD and could serve as a reference for the interpretation of disease activity biomarkers in this highly understudied population. References: [1] Dallmann, A. et al. Gestation-Specific Changes in the Anatomy and Physiology of Healthy Pregnant Women: An Extended Repository of Model Parameters for Physiologically Based Pharmacokinetic Modeling in Pregnancy. Clin. Pharmacokinet. 56, 1303–1330 (2017). [2] Abduljalil, K., Furness, P., Johnson, T. N., Rostami-Hodjegan, A. & Soltani, H. Anatomical, Physiological and Metabolic Changes with Gestational Age during Normal Pregnancy: A Database for Parameters Required in Physiologically Based Pharmacokinetic Modelling. Clin. Pharmacokinet. 51, 365–396 (2012). [3] Eke, A. C. An update on the physiologic changes during pregnancy and their impact on drug pharmacokinetics and pharmacogenomics. J. Basic Clin. Physiol. Pharmacol. 33, 581–598 (2022). [4] Gremese, E. et al. Serum Albumin Levels: A Biomarker to Be Repurposed in Different Disease Settings in Clinical Practice. J. Clin. Med. 12, 6017 (2023). [5] Pepys, M. B. & Hirschfield, G. M. C-reactive protein: a critical update. J. Clin. Invest. 111, 1805–1812 (2003). [6] Mahadevan, U. et al. Global Consensus Statement on the Management of Pregnancy in Inflammatory Bowel Disease. Clin. Gastroenterol. Hepatol. Off. Clin. Pract. J. Am. Gastroenterol. Assoc. 23, S1–S60 (2025). [7] Duvnjak, Z. et al. Physiologically Motivated Sequential Population Modeling of Albumin Trends and Vedolizumab Pharmacokinetics for Pregnancy Dosing Regimen Optimization. Clin. Pharmacol. Ther. 119, 457–469 (2026). [8] Julsgaard, M. et al. Vedolizumab clearance in neonates, susceptibility to infections and developmental milestones: a prospective multicentre population-based cohort study. Aliment. Pharmacol. Ther. 54, 1320–1329 (2021). [9] Julsgaard, M. et al. Infant Ustekinumab Clearance, Risk of Infection, and Development After Exposure During Pregnancy. Clin. Gastroenterol. Hepatol. Off. Clin. Pract. J. Am. Gastroenterol. Assoc. 23, 134–143 (2025).
Reference: PAGE 34 (2026) Abstr 12268 [www.page-meeting.org/?abstract=12268]
Poster: Drug/Disease Modelling - Other Topics