Addition of Vitamin A Intake Data during Compartmental Modeling of Retinol Kinetics in Theoretical Humans Leads to Accurate Prediction of Vitamin A Total Body Stores and Kinetic Parameters in Studies of Reasonable Duration

Use of Population-Based Compartmental Modeling and Retinol Isotope Dilution to Study Vitamin A Kinetics and Total Body Stores among Ghanaian Women of Reproductive Age

Background

Limited data are available on vitamin A kinetics and total body stores (TBS) in women. Such information can be obtained using compartmental modeling and retinol isotope dilution (RID).

Objectives

Objectives were to apply population-based ("super-subject") modeling to determine retinol kinetics in nonpregnant Ghanaian women of reproductive age and to use RID to predict TBS in the group and its individuals.

Methods

Women (n = 89) ingested a dose of [2H6]retinyl acetate and blood samples (3/woman) were collected from 6 h to 91 d, with all participants sampled at 14 d, about half at either 21 or 28 d, and each at one other time. Composite data (plasma retinol fraction of dose; FDp) were analyzed using Simulation, Analysis and Modeling software to obtain kinetic parameters, TBS, and other state variables as well as model-derived values for the RID composite coefficient FaS. The latter were used in the RID equation TBS (μmol) = FaS × 1/SAp (where SAp is plasma retinol specific activity) to predict TBS at various times.

Results

Model-predicted TBS was 973 μmol (n = 87). Geometric mean RID-predicted TBS was 965, 926, and 1006 μmol at 14, 21, and 28 d, respectively, with wide ranges [for example, 252-3848 μmol on day 14 (n = 86)]; TBS predictions were similar at later times. Participants had a mean 2 y of vitamin A in stores and estimated liver vitamin A concentrations in the normal range. Model-predicted vitamin A disposal rate was 1.3 μmol/d and plasma recycling number was 37.

Conclusions

Super-subject modeling provides an estimate of group mean TBS as well as group-specific values for the RID coefficient FaS; the latter can be used to confidently predict TBS by RID for individual participants in the group under study or in similar individuals at 14 d or more after isotope ingestion.

Trial registration number

Trial is registered (NCT04632771) at https://clinicaltrials.gov.

Use of Theoretical Women and Model-Based Compartmental Analysis to Evaluate the Impact of Vitamin A Intake with or without a Daily Vitamin A Supplement on Vitamin A Total Body Stores and Balance During Lactation

BACKGROUND

Inadequate vitamin A (VA) intake is common among lactating women in many communities worldwide, but high-dose VA supplementation for postpartum women is not recommended by the World Health Organization as an effective intervention.

OBJECTIVES

To simulate the impact of VA intake via diet and daily VA supplements on VA total body stores (TBS) and balance in theoretical lactating women with low/moderate TBS.

METHODS

We studied 6 theoretical subjects with assigned values for TBS from 219-624 μmol. Using Simulation, Analysis, and Modeling software and a previously published compartmental model for whole-body VA metabolism, we simulated TBS over 6 mo of established lactation for each subject under 4 conditions: 1) prelactation VA intake was increased to maintain VA balance (LSS); 2) prelactation VA intake was maintained (NLSS); 3) VA intake was the same as 2) but a daily VA supplement (2.8 μmol/d) was added (NLSS+S); and 4) VA intake was as 1) and the daily VA supplement was included (LSS+S).

RESULTS

To compensate for the loss of VA via milk while VA balance was maintained (LSS) over 6 mo of lactation, VA intake had to increase by 0.8-1.87 μmol/d (n = 6) compared with NLSS. Over 6 mo of NLSS treatment, VA balance was negative (geometric mean, -0.77 μmol/d) compared with LSS, whereas balance was positive under NLSS+S and LSS+S conditions (0.75 and 1.5 μmol/d, respectively). For LSS, the proportion of total VA disposal was 37% via breastmilk, 32% from VA stores, and 32% from nonstorage tissues.

CONCLUSIONS

Adding a daily VA supplement (2.8 μmol/d) to the diet of lactating women with suboptimal VA intake may effectively counterbalance the negative VA balance resulting from the output of VA via breastmilk and thus benefit both mother and infant by maintaining or increasing VA stores and breastmilk VA concentration.

Use of Compartmental Modeling and Retinol Isotope Dilution to Determine Vitamin A Stores in Young People with Sickle Cell Disease Before and After Vitamin A Supplementation

BACKGROUND

Suboptimal plasma retinol concentrations have been documented in US children with sickle cell disease (SCD) hemoglobin SS type (SCD-HbSS), but little is known about vitamin A kinetics and stores in SCD.

OBJECTIVES

The objectives were to quantify vitamin A total body stores (TBS) and whole-body retinol kinetics in young people with SCD-HbSS and use retinol isotope dilution (RID) to predict TBS in SCD-HbSS and healthy peers as well as after vitamin A supplementation in SCD-HbSS subjects.

METHODS

Composite plasma [13C10]retinol response data collected from 22 subjects with SCD-HbSS for 28 d after isotope ingestion were analyzed using population-based compartmental modeling ("super-subject" approach); TBS and retinol kinetics were quantified for the group. TBS was also calculated for the same individuals using RID, as well as for healthy peers (n = 20) and for the subjects with SCD-HbSS after 8 wk of daily vitamin A supplements (3.15 or 6.29 μmol retinol/d [900 or 1800 μg retinol activity equivalents/d]).

RESULTS

Model-predicted group mean TBS for subjects with SCD-HbSS was 428 μmol, equivalent to ∼11 mo of stored vitamin A; vitamin A disposal rate was 1.3 μmol/d. Model-predicted TBS was similar to that predicted by RID at 3 d postdosing (mean, 389 μmol; ∼0.3 μmol/g liver); TBS predictions at 3 compared with 28 d were not significantly different. Mean TBS in healthy peers was similar (406 μmol). RID-predicted TBS for subjects with SCD-HbSS was not significantly affected by vitamin A supplementation at either dose.

CONCLUSIONS

Despite differences in plasma retinol concentrations, TBS was the same in subjects with SCD-HbSS compared with healthy peers. Because 56 d of vitamin A supplementation at levels 1.2 to 2.6 times the Recommended Dietary Allowance did not increase TBS in these subjects with SCD-HbSS, further work will be needed to understand the effects of SCD on retinol metabolism. This trial was registered as NCT03632876 at clinicaltrials.gov.

Direct Use of Plasma or Milk Vitamin A Specific Activity Data to Model Retinol Kinetics and Predict Vitamin A Stores in Theoretical Lactating Women

BACKGROUND

Many applications of the Simulation, Analysis and Modeling software use data on the fraction of an orally administered tracer dose (FD) in plasma; thus, researchers must scale-up measured analyte concentration to the total plasma pool. For studies in lactating women, estimating breast milk pool size is challenging.

OBJECTIVES

The objectives were to determine whether the standard vitamin A modeling approach using FD data could be modified to use vitamin A specific activity in milk (SAm) and/or plasma (SAp) for compartmental analysis of vitamin A kinetics and status in theoretical lactating women.

METHODS

Using 12 previously studied theoretical subjects with a wide range of assigned values for vitamin A total body stores (TBS) and the coefficient ("FaS") needed to predict TBS using a retinol isotope dilution equation, we simulated data for SAp and SAm for 49 d after oral administration of labeled vitamin A. Then we modeled datasets for SAp and SAm, as well as only SAp or SAm, incorporating a linear scaling factor to automatically convert SA to FD and including several physiologically reasonable constraints as input data. As outcomes, we compared model-predicted TBS and FaS to assigned values.

RESULTS

Scaling factors effectively adjusted SA data to adequately predict vitamin A mass in plasma and breast milk pools. Data for SAp and SAm provided model predictions of TBS that were comparable to assigned values (range: 85-107%); using only SAp, ratios ranged from 92% to 108% and for SAm from 85% to 108%. Parallel results were obtained for simulated FaS.

CONCLUSIONS

Results show that SA data from plasma and/or milk can be used directly for modeling vitamin A during lactation in theoretical subjects, providing accurate estimates of TBS and FaS. Results suggest that, in free-living lactating women, researchers might measure only SAp or only SAm and adequately describe whole-body vitamin A metabolism and status.

Compartmental Modeling of Vitamin A Stable Isotope Data from Milk or Plasma Provides Comparable Predictions of Vitamin A Stores in Theoretical Lactating Women

BACKGROUND

Previous compartmental models describing and quantifying whole-body vitamin A (VA) metabolism have been developed from plasma retinol kinetic data after human subjects ingest stable isotope-labeled VA. For humans, models based on data obtained from other sampling sites (e.g., excreta or milk) have not been proposed.

OBJECTIVES

Our objective was to determine whether comparable model predictions of VA total body stores (TBS) in theoretical lactating women were obtained using tracer data from only retinol in plasma or VA in milk.

METHODS

We used Simulation, Analysis and Modeling software to simulate values for TBS and the coefficients used in the retinol isotope dilution (RID) equation TBS = FaS/SAp (Fa, fraction of dose in stores; S, retinol specific activity (SA) in plasma/SA in stores; SAp, specific activity in plasma). We compared individual subject predictions of TBS and FaS based on modeling only plasma or only milk tracer data to previous results ("assigned values") for 12 theoretical lactating women when modeling was done based on tracer data for chylomicron retinyl esters, plasma retinol, and milk VA.

RESULTS

For subjects with a wide range of TBS, model-predicted TBS based on only plasma data were comparable with assigned values (range: 94%-106%). Using only milk data, predictions ranged from 72% to 178%, but when VA intake was included in modeling, predictions were improved (97%-102%). Similar results were obtained for simulated FaS.

CONCLUSIONS

If confirmed in free-living lactating women, results indicate that, similar to models based on serial plasma sampling, a model for whole-body VA kinetics, including predictions of TBS and FaS, can be identified based on tracer data for VA in milk when VA intake is included as a modeling constraint. Milk data have not been previously used for compartmental modeling of VA in humans.

Development of a Compartmental Model for Studying Vitamin A Kinetics and Status in Theoretical Lactating Women

BACKGROUND

Low vitamin A status and suboptimal milk vitamin A concentrations are problems in many populations worldwide. However, limited research has been done on whole-body vitamin A kinetics in women of reproductive age, especially during lactation.

OBJECTIVES

Goals were to develop compartmental models describing retinol kinetics in theoretical nonlactating (NL) and lactating (L) women and to determine whether the retinol isotope dilution (RID) method accurately predicted vitamin A total body stores (TBS) in the groups and individuals.

METHODS

We adapted 12 previously-used theoretical females with assigned values for retinol kinetic parameters and TBS (225-1348 μmol); subjects were NL or L (nursing one 3- to 6-mo-old infant) during 49-d kinetic studies after isotope dosing. We used an established compartmental model, adding a compartment for chylomicrons and, for L, another for mammary gland milk with inputs from holo-retinol-binding protein and chylomicron retinyl esters and output to milk. Using compartmental analysis, we simulated tracer responses in compartments of interest and calculated TBS using the RID equation TBS =   FaS/SAp [Fa, fraction of dose in stores; S, retinol specific activity in plasma/specific activity in stores; SAp, specific activity of retinol in plasma].

RESULTS

Models for both groups were well identified. Simulated plasma tracer responses were similar for NL and L, with L always below NL; milk tracer paralleled plasma from 10 d postdosing. Geometric mean FaS ratios (L/NL) were ∼0.75 during days 2-30. Using appropriate group FaS, RID provided accurate TBS predictions for >80% of NL and L subjects after day 18 when CV% for FaS was ∼10%.

CONCLUSIONS

These new physiologically-based models for vitamin A kinetics may be useful for future research in women of reproductive age. Results indicate that, in groups like these, RID to assess an individual's vitamin A status should be done at 21-28 d after isotope dosing.

The Use of Datasets for Theoretical Subjects to Validate Vitamin A-Related Methods and Experimental Designs

We review recent work in which model-based compartmental analysis has been applied to data for theoretical human subjects in order to study questions related to vitamin A kinetics and metabolism. Using model simulations in this way, one can validate experimental designs, evaluate or improve vitamin A assessment methods, study the influence of perturbations on assessment methods, and/or advance information related to retinol kinetics. We also provide some information on the rationale for assigning physiologically appropriate values for specified characteristics [e.g., plasma retinol concentration, vitamin A total body stores (TBS), vitamin A intake] to hypothetical individuals, and in addition, we outline how one might first select an appropriate compartmental model for whole-body vitamin A metabolism and then specify physiologically reasonable values for the associated retinol kinetic parameters. In the studies discussed here, the Simulation, Analysis, and Modeling software was used to simulate responses in key model compartments for hypothetical subjects so that model predictions could be compared to assigned values or projected outcomes. For example, in the case of the retinol isotope dilution (RID) method that is used to assess vitamin A status, application of this approach has provided a way to evaluate the accuracy of TBS predictions under different steady state and non-steady state conditions, thus increasing confidence about the validity of RID results obtained in the field. Although datasets for theoretical subjects have been used to evaluate protocols in pharmacokinetics, to our knowledge, other nutrition researchers have not previously used approaches such as those described here. Our results to date indicate that this strategy has the potential to provide useful information related not only to vitamin A but perhaps to other nutrients as well.

Influence of Vitamin A Status on the Choice of Sampling Time for Application of the Retinol Isotope Dilution Method in Theoretical Children

BACKGROUND

Vitamin A status may influence the choice of a blood sampling time for applying the retinol isotope dilution (RID) equation to predict vitamin A total body stores (TBS) in children.

OBJECTIVES

We aimed to identify time(s) after administration of labeled vitamin A that provide accurate estimates of TBS in theoretical children with low or high TBS.

METHODS

We postulated 2- to 5-y-old children (12/group) with low (<200 μmol) or high TBS (≥700 μmol) and used compartmental analysis to simulate individual subject values for the RID equation TBS =   FaS/SAp (Fa, fraction of dose in stores; S, retinol specific activity in plasma/in stores; SAp, retinol specific activity in plasma). Using individual SAp and group geometric mean FaS values from 1-28 d, we calculated individual and group mean TBS and compared them to assigned values.

RESULTS

Mean TBS was accurately predicted for both groups at all times. For individuals, predicted and assigned TBS were closest when the CV% for FaS was low [12-14%; 4-13 d (low), 12-28 d (high)]. The mean percentage error for TBS was <10% from 2-19 d (low) and 7-28 d (high). Predicted TBS was within 25% of assigned TBS for ≥80% of children from 3-23 d (low) and 9-28 d (high). Within groups, RID tended to overestimate lower TBS and underestimate higher TBS.

CONCLUSIONS

Using a good estimate for FaS, accurate RID predictions of TBS for individuals will be obtained at many times. If vitamin A status is low, results indicate that early sampling (e.g., 4-13 d) is optimal; if vitamin A status is high, sampling at 12-28 d is indicated. When vitamin A status is unknown, sampling at 14 d is recommended, or a super-subject design can be used to obtain the group mean FaS at various times for RID prediction of TBS in individuals.

Pregnancy and Lactation Alter Vitamin A Metabolism and Kinetics in Rats under Vitamin A-Adequate Dietary Conditions

BACKGROUND

Vitamin A (VA) plays critical roles in prenatal and postnatal development; however, limited information is available regarding maternal VA metabolism during pregnancy and lactation.

OBJECTIVES

We investigated the impact of pregnancy and lactation on VA metabolism and kinetics in rats, hypothesizing that changes in physiological status would naturally perturb whole-body VA kinetics.

METHODS

Eight-week old female rats (n = 10) fed an AIN-93G diet received an oral tracer dose of 3H-labeled retinol to initiate the kinetic study. On d 21 after dosing, six female rats were mated. Serial blood samples were collected from each female rat at selected times after dose administration until d 14 of lactation. Model-based compartmental analysis was applied to the plasma tracer data to develop VA kinetic models.

RESULTS

Our compartmental model revealed that pregnancy resulted in a gradual increase in hepatic VA mobilization, presumably to support different stages of fetal development. Additionally, the model indicates that during lactation, VA derived from dietary intake was the primary source of VA delivered to the mammary gland for milk VA secretion.

CONCLUSION

During pregnancy and lactation in rats with an adequate VA intake and previous VA storage, the internal redistribution of VA and increased uptake from diet supported the maintenance of VA homeostasis.

Use of Model-Based Compartmental Analysis and Theoretical Data to Further Explore Choice of Sampling Time for Assessing Vitamin A Status in Groups and Individual Human Subjects by the Retinol Isotope Dilution Method

BACKGROUND

An optimal blood sampling time for application of the retinol isotope dilution (RID) method for predicting vitamin A total body stores (TBS) (i.e., vitamin A status) has not been established.

OBJECTIVES

Objectives were to identify sampling times that provide accurate estimates of TBS by RID in groups and individuals by applying compartmental modeling to data for theoretical adults and children.

METHODS

We selected previously generated hypothetical adults and children (20 per group) that had a wide range of assigned values for TBS and vitamin A kinetic parameters. We used the Simulation, Analysis and Modeling software to simulate individual kinetic responses; then we calculated geometric mean values for the RID equation coefficients and each individual's plasma retinol specific activity at various times, using those values to predict group mean and individual subject TBS. Predicted values for TBS were compared with assigned values.

RESULTS

Accurate estimates of group mean TBS were obtained at all sampling times from 1 to 30 d in both adults and children. For individuals, correlations between RID-predicted TBS and assigned values increased with time in the adults (R2 = 0.80 at day 14, 0.96 at day 21, and 0.99 at day 28); a similar trend was observed for the children, with R2 = 0.82 at day 7 and increasing to 0.97 at days 21 and 28 (P < 0.001 for all comparisons).

CONCLUSIONS

Although no single, unique time provided the most accurate prediction of TBS for all individuals within these groups, applying the RID method at 21 or 28 d yielded predictions that were within 25% of assigned values for 90% or 95% of adults, respectively; corresponding values for children were 80% from 10 to 20 d, and 85% at 21 and 28 d. For most subjects, early times (<14 d for adults and <10 d for children) provided less accurate predictions.

Development of a Compartmental Model to Investigate the Influence of Inflammation on Predictions of Vitamin A Total Body Stores by Retinol Isotope Dilution in Theoretical Humans

BACKGROUND

Inflammation, both acute and chronic, is associated with reductions in the synthesis of retinol-binding protein (RBP) and the concentration of retinol in plasma. Consequently, it is currently recommended that the retinol isotope dilution (RID) method not be used to estimate vitamin A total body stores (TBS) in subjects with inflammation.

OBJECTIVES

To apply compartmental analysis to study the effects of inflammation on hepatic apo-RBP input, plasma retinol pool size, and RID-predicted TBS in theoretical subjects with known steady state values for these parameters.

METHODS

We selected 6 previously generated hypothetical subjects who ingested a dose of stable isotope-labeled vitamin A (day 0). Starting with each subject's published steady state model for retinol tracer kinetics, we developed a parallel model for unlabeled retinol and RBP that incorporated links between these entities and tied liver retinol secretion to RBP availability. Then we perturbed the steady state model by initiating chronic or acute inflammation on day 0 or acute inflammation on day 3 or 9 and simulating results for RBP, plasma retinol, and TBS.

RESULTS

Chronic inflammation led to substantial reductions in RID-predicted TBS for at least 2 weeks after tracer administration. In contrast, acute inflammation induced on day 0 or 3 resulted in less dramatic impacts on TBS (37% or 20% reduction, respectively, from steady state levels, with levels rebounding by 14 days). When inflammation was induced 9 days after administration of the tracer, the effects on predicted TBS were minimal. Overall, for acute inflammation initiated at 0, 3, or 9 days, accurate predictions of TBS were obtained by 2 weeks.

CONCLUSIONS

Compartmental analysis can be applied in the novel way described here to study the influence of perturbations such as inflammation on predictions of vitamin A status using RID. Such an approach has potential value for studying other perturbations and different nutrients.

Vitamin A Absorption Efficiency Determined by Compartmental Analysis of Postprandial Plasma Retinyl Ester Kinetics in Theoretical Humans

BACKGROUND

Better methods are needed for determining vitamin A absorption efficiency in humans to support development of dietary recommendations and to improve the accuracy of predictions of vitamin A status.

OBJECTIVES

We developed and evaluated a method for estimating vitamin A absorption efficiency based on compartmental modeling of theoretical data on postprandial plasma retinyl ester (RE) kinetics.

METHODS

We generated data on plasma RE and retinol kinetics (30 min to 8 h or 56 d, respectively) after oral administration of labeled vitamin A for 12 theoretical adults with a range of values assigned for vitamin A absorption (55-90%); we modeled all data to obtain best-fit values for absorption and other parameters using Simulation, Analysis, and Modeling software. We then modeled RE data only (16 or 10 samples), with or without added random error, and compared assigned to predicted absorption values. We also compared assigned values to areas under RE response curves (RE AUCs).

RESULTS

We confirmed that a unique value for vitamin A absorption cannot be identified by modeling plasma retinol tracer kinetics. However, when RE data were modeled, predicted vitamin A absorptions were within 1% of assigned values using data without error and within 12% when 5% error was included. When the sample number was reduced, predictions were still within 13% for 10 of the 12 subjects and within 23% overall. Assigned values for absorption were not correlated with RE AUC (P = 0.21).

CONCLUSIONS

We describe a feasible and accurate method for determining vitamin A absorption efficiency that is based on compartmental modeling of plasma RE kinetic data collected for 8 h after a test meal. This approach can be used in a clinical setting after fasting subjects consume a fat-containing breakfast meal with a known amount of vitamin A or a stable isotope label.

Vitamin A Absorption Determined in Rats Using a Plasma Isotope Ratio Method

BACKGROUND

Better methods are needed for determining vitamin A absorption efficiency.

OBJECTIVE

Our objective was to measure vitamin A absorption in rats by adapting a plasma isotope ratio method previously used to determine cholesterol absorption.

METHODS

Male Sprague-Dawley rats [n = 14; 340 ± 16 g (mean ± SD)] received an oral tracer dose of [3H]retinyl acetate in oil plus an intravenous dose of [14C]vitamin A-labeled lymph prepared in a donor rat that had received [14C]retinyl acetate intraduodenally. Blood samples were collected on days 1, 2, 3, 6, 9, and 12, and plasma was analyzed for 3H and 14C; vitamin A absorption was calculated for each sample as (fraction of oral dose/fraction of intravenous dose) × 100. Radioactivity was also measured in feces and urine collected as pools on days 3, 6, 9, and 12 and in liver and remaining carcass on day 12.

RESULTS

Vitamin A absorption calculated as the plasma isotope ratio was >100% on day 1, 78% ± 5% on day 6, 76% ± 5% on day 9, and 74% ± 5% on day 12; fitting the data to an exponential function plus a constant predicted an absorption of 75% by day 14. Recovery of the oral dose in feces (day 0 to day 6) was low (6.2% ± 0.84%, n = 10) and the mean isotope ratio in day 9-12 urine pool was lower than that in plasma.

CONCLUSIONS

The plasma isotope ratio holds promise for estimating vitamin A absorption, but additional work is needed to determine how long studies need to be and if the doses should be administered simultaneously. For application of this method in humans, artificial chylomicrons labeled with a stable isotope of retinyl acetate could be used for the intravenous dose, with a different isotope required for the oral dose.

The "Super-Child" Approach Is Applied To Estimate Retinol Kinetics and Vitamin A Total Body Stores in Mexican Preschoolers

BACKGROUND

Retinol isotope dilution (RID) and model-based compartmental analysis are recognized techniques for assessing vitamin A (VA) status. Recent studies have shown that RID predictions of VA total body stores (TBS) can be improved by using modeling and that VA kinetics and TBS in children can be effectively studied by applying population modeling ("super-child" approach) to a composite data set.

OBJECTIVES

The objectives were to model whole-body retinol kinetics and predict VA TBS in a group of Mexican preschoolers using the super-child approach and to use model predictions of RID coefficients to estimate TBS by RID in individuals.

METHODS

Twenty-four healthy Mexican children (aged 3-6 y) received an oral dose (2.96 μmol) of [13C10]retinyl acetate in corn oil. Blood samples were collected from 8 h to 21 d after dosing, with each child sampled at 4 d and at 1 other time. Composite data for plasma labeled retinol compared with time were analyzed using a 6-component model to obtain group retinol kinetic parameters and pool sizes. Model-predicted TBS was compared with mean RID predictions at 4 d; RID estimates at 4 d were compared with those calculated at 7-21 d.

RESULTS

Model-predicted TBS was 1097 μmol, equivalent to ∼2.4 y-worth of VA; using model-derived coefficients, group mean RID-predicted TBS was 1096 μmol (IQR: 836-1492 μmol). TBS at 4 d compared with a later time was similar (P = 0.33). The model predicted that retinol spent 1.5 h in plasma during each transit and recycled to plasma 13 times before utilization.

CONCLUSIONS

The super-child modeling approach provides information on whole-body VA kinetics and can be used with RID to estimate TBS at any time between 4 and 21 d postdose. The high TBS predicted for these children suggests positive VA balance, likely due to large-dose VA supplements, and warrants further investigation.

Better Predictions of Vitamin A Total Body Stores by the Retinol Isotope Dilution Method Are Possible with Deeper Understanding of the Mathematics and by Applying Compartmental Modeling

Retinol isotope dilution (RID) is a well-accepted technique for assessing vitamin A status [i.e., total body stores (TBS)]. Here, in an effort to increase understanding of the method, we briefly review RID equations and discuss their included variables and their coefficients (i.e., assumptions that account for the efficiency of absorption of an orally administered tracer dose of vitamin A, mixing of the dose with endogenous vitamin A, and loss due to utilization). Then, we focus on contributions of another technique, model-based compartmental analysis and especially the "super-person" approach, that advance the RID method. Specifically, we explain how adding this modeling component, which involves taking 1 additional blood sample from each subject, provides population-specific estimates for the RID coefficients that can be used in the equation instead of values derived from the literature; using model-derived RID coefficients results in improved confidence in predictions of TBS for both a group and its individuals. We note that work is still needed to identify the optimal time for applying RID in different groups and to quantify vitamin A absorption efficiency. Finally, we mention other contributions of modeling, including the use of theoretical data to verify the accuracy of RID predictions and the additional knowledge that model-based compartmental analysis provides about whole-body vitamin A kinetics.

Use of Model-Based Compartmental Analysis and a Super-Child Design to Study Whole-Body Retinol Kinetics and Vitamin A Total Body Stores in Children from 3 Lower-Income Countries

BACKGROUND

Model-based compartmental analysis has been used to describe and quantify whole-body vitamin A metabolism and estimate total body stores (TBS) in animals and humans.

OBJECTIVES

We applied compartmental modeling and a super-child design to estimate retinol kinetic parameters and TBS for young children in Bangladesh, Guatemala, and the Philippines.

METHODS

Children ingested [13C10]retinyl acetate and 1 or 2 blood samples were collected from each child from 6 h to 28 d after dosing. Temporal data for fraction of dose in plasma [13C10]retinol were modeled using WinSAAM software and a 6-component model with vitamin A intake included as weighted data.

RESULTS

Model-predicted TBS was 198, 533, and 1062 μmol for the Bangladeshi (age, 9-17 mo), Filipino (12-18 mo), and Guatemalan children (35-65 mo). Retinol kinetics were similar for Filipino and Guatemalan groups and generally faster for Bangladeshi children, although fractional transfer of plasma retinol to a larger exchangeable storage pool was the same for the 3 groups. Recycling to plasma from that pool was ∼2.5 times faster in the Bangladeshi children compared with the other groups and the recycling number was 2-3 times greater. Differences in kinetics between groups are likely related to differences in vitamin A stores and intakes (geometric means: 352, 727, and 764 μg retinol activity equivalents/d for the Bangladeshi, Filipino, and Guatemalan children, respectively).

CONCLUSIONS

By collecting 1 or 2 blood samples from each child to generate a composite plasma tracer data set with a minimum of 5 children/time, group TBS and retinol kinetics can be estimated in children by compartmental analysis; inclusion of vitamin A intake data increases confidence in model predictions. The super-child modeling approach is an effective technique for comparing vitamin A status among children from different populations. These trials were registered at www.clinicaltrials.gov as NCT03000543 (Bangladesh), NCT03345147 (Guatemala), and NCT03030339 (Philippines).