(P01-007-23) Use of Population Modeling and Retinol Isotope Dilution to Study Vitamin A Kinetics and Total Body Stores in Ghanaian Women of Reproductive Age

Objectives: Objectives were to use model-based compartmental analysis with a super-subject design and retinol isotope dilution (RID) to study vitamin A (VA) kinetics and VA total body stores (TBS) among women in Northern Ghana and to compare TBS predicted by RID at multiple sampling times.

Methods: Non-pregnant, non-lactating women (15-49 y of age; n=98) ingested a tracer dose of [²H₆]retinyl acetate (6.9 µmol retinol equivalent) on d 0. Blood (6 mL) was collected from all women on d 14, from half on d 21, and from the other half on d 28. A third sample was collected at one of 13 other times (6, 9, or 12 h or 1, 2, 4, 7, 11, 35, 49, 63, 77, or 91 d; ~7 women/time). Serum was analyzed by LC-MS/MS to determine retinol specific activity (SA_p = fraction of dose/µmol). A composite (super-subject) dataset of geometric mean (GM) SA_p vs. time was fit to a previously published compartmental model (Green MH, Green JB. J Nutr 2021;151:2068–2074) using Simulation, Analysis and Modeling (SAAM) software to obtain VA kinetic parameters and TBS for the group. Individual subject TBS were predicted by the RID equation TBS = FaS x (1/SA_p) at 8 sampling times between 14 and 91 d, using individual SA_p and values for the composite coefficient FaS calculated from the model; then GM TBS was calculated at each time.

Results: Model-predicted TBS for the group was 911 µmol, equivalent to 1.9 y of VA stores; model-derived values for FaS were 0.881 (d 14), 0.737 (d 21) and 0.688 (d 28). RID-predicted GM TBS at d 14 was 872 µmol (min, max: 226, 3448 µmol; n=85) and values did not differ among women sampled on d 14 and d 21 (871 vs. 882 µmol; n=39; p=0.24) or d 14 and d 28 (857 vs. 890 µmol; n=12; p=0.52). Over all sampling times from d 14-91, predicted GM TBS ranged from 872-950 µmol. Assuming 80% of TBS is in liver and liver weight is 2.4% of body weight, no women had estimated liver VA concentration < 0.1 µmol/g or >3.0 µmol/g; the GM was 0.517 µmol/g on d 14.

Conclusions: This is the first super-subject VA kinetic study in which modeling was done using SA_p data and RID predictions were systematically evaluated at different times. Results show that these Ghanaian women had adequate VA status, with a wide range of TBS. We conclude that RID provides comparable predictions of TBS from 14-91 d postdosing in women with adequate VA status.

Funding Sources: Helen Keller International, Bill & Melinda Gates Foundation