Ocampo-Pelland, A.S. (1,3), Gastonguay, M.R. (1,2,3), French, J.L.(2,3), Riggs, M.M. (2)
(1) Department of Biomedical Engineering, University of Connecticut, USA, (2)Metrum Research Group, LLC, Tariffville, Connecticut, USA (3)Metrum Institute, Tariffville, Connecticut, USA
Objectives: Association of Vitamin D3 (D3) and its metabolite (25OHD3) exposure with various diseases, including bone health, diabetes, and cancer, has become an active area of research[1]. Clinical studies investigating these relationships with D3 and 25OHD3 vary in dosing regimen, assays, demographics, and control of endogenous D3 production. This leads to uncertain and conflicting exposure-related associations with D3 and 25OHD3. To elucidate this parent-metabolite system, a population PK (PPK) model was developed to predict D3 and 25OHD3 from varied doses and administration routes. Sources of variability related to 25OHD3 baseline (BL), weight (WT), and assay type were explored.
Methods: Public source PK data pertaining to D3 and 25OHD3 in healthy or osteoporotic populations, including 57 studies representing 5406 individuals (25 individual-level and 122 group-level units), were selected using specified search criteria in PUBMED. Data included IV, oral, single and multiple dose data: dose ranges for D3 (400-100000 IU/d) and 25OHD3 (15-1000 ug/d). A nonlinear (NL) mixed effects model was developed to simultaneously model the D3 and 25OHD3 PK (NONMEM v7.2). Model development explored 1- and 2-compartment (CMT) models with first-order absorption rate constant (ka) and linear or NL clearance (CLNL). Unit-level random effects and residual errors were weighted by arm sample size[2].
Results: D3 and 25OHD3 dispositions were described by 2-CMT models with NL and linear clearances (CL), respectively. D3 model estimates, apparent to oral administration were: CLNL (VMAX=1.62 nmol/h, KM=6.39 nmol/L), central volume (VC=15.5 L), intercompartmental clearance (Q=0.185 L/h), peripheral volume (VP=2333 L/h), BL concentration (3.75 nmol/L), and mean endogenous production rate (220 IU/d). For 25OHD3: CLM=0.0153 L/h, VCM=4.35 L, VTM=6.87 L, QM=0.0507 L/h. The same ka for D3 and 25OHD3 was assumed (0.323h-1); CLNL was set equal to 25OHD3 formation rate. Simulations showed an inverse relationship between 25OHD3 and BL 25OHD3. HPLC-MS was the reference assay for 25OHD3; RIA was most similar but differences in precision and bias were estimated for competitive protein binding assay and chemiluminescence. WT was too inconsistently reported in publications to discern its covariate effect in the PPK.
Conclusions: The PK of D3 and 25OHD3 suggest that estimation of CLNL was important when considering comparisons of D3 and 25OHD3 exposure across studies and dosing regimens.
References:
[1] T. Hagenau, R. Vest, T. N. Gissel, C. S. Poulsen, M. Erlandsen, L. Mosekilde, and P. Vestergaard, “Global vitamin D levels in relation to age, gender, skin pigmentation and latitude: An ecologic meta-regression analysis,” Osteoporos. Int., vol. 20, pp. 133–140, 2009.
[2] J. E. Ahn and J. L. French, “Longitudinal aggregate data model-based meta-analysis with NONMEM: Approaches to handling within treatment arm correlation,” J. Pharmacokinet. Pharmacodyn., vol. 37, pp. 179–201, 2010.
Reference: PAGE 24 () Abstr 3342 [www.page-meeting.org/?abstract=3342]
Poster: Drug/Disease modeling - Endocrine