Development of a pregnancy-specific physiologically based pharmacokinetic model for cadmium and estimation of fetal internal dose using serial maternal biomarkers and clinical chemistry parameters.

Cadmium (Cd) poses developmental risks during pregnancy, yet single time-point biomarkers inadequately capture cumulative fetal internal dose. We developed a pregnancy-specific semi-mechanistic physiologically based pharmacokinetic (PBPK) model to estimate fetal Cd exposure trajectories in 2778 mother-infant pairs from the Korea Children's Environmental Health Study (Ko-CHENS). The ten-compartment maternal-fetal model incorporated serial blood Cd at early (T1; gestational weeks 12-16) and late pregnancy (T3; weeks 28-34) with paired cord blood Cd, and was individually parameterized using four clinical chemistry variables (serum albumin, calcium, blood urea nitrogen, and hs-CRP), enabling person-specific modulation of gastrointestinal absorption and placental transfer kinetics. Calibration yielded a mean fold error of 0.999 for cord blood Cd, with 87.5% of T3 predictions within 2-fold of observed values. External validation against second-trimester urinary Cd from the independent LIFE-EPIC cohort (n = 200) confirmed transferability (GM ratio = 1.05; MPE = +5.0%). Cord blood Cd was invariant across T1 quintiles (ANOVA p = 0.53, η2 < 0.001); the empirical f/m power-law exponent (β = -1.00, r2 = 0.999) established independently of the PBPK model confirms cord blood is mechanistically unsuited as an individual-level fetal exposure metric, consistent with the PBPK-estimated exponent (β = -0.91) reflecting saturable metallothionein-mediated placental sequestration. Despite a 3.8-fold range in maternal T1 Cd, PBPK-derived fetal AUC varied by only 3.3%. Maternal age attenuated placental transfer efficiency, and clinical chemistry parameterization improved cord blood Cd prediction (MPE: +12.2% → +7.2%). PBPK-derived dose metrics provide superior hazard characterization for prenatal Cd risk assessment.