Abstract
At present there is little information regarding the quantitative vitamin A value of specific foods or meal patterns. Such information would aid in a more precise determination of the suitability of certain foods or diets to meet minimum and optimal nutritional needs for vitamin A. This knowledge deficit stems largely from the lack of well-developed models to assess the bioavailability of carotenoids, particularly provitamin A carotenoids, from foods. Literature values of carotenoid bioavailability (percent of ingested carotene absorbed) range from 1% to 99%, and variability is often high both within and between treatments. Current models generally fall into two categories: those employing outcomes that reflect efficacy, i.e., that rely on changes in one or more indicators of vitamin A status; and those which provide a more direct estimate of absorption and/or conversion efficiency of single doses of provitamin A carotenoids. the latter include oral–faecal balance models and post-prandial chylomicron retinyl ester response models. Absolute bioavailability, i.e., the mass of newly absorbed vitamin A derived from a given mass of provitamin A carotenoid consumed, is inherently more problematic to assess than relative bioavailability, in which two or more treatments are compared. One missing and key piece of information is the extent of post-absorptive conversion of provitamin A carotenoids to vitamin A in humans. Without such information, the contribution of absorbed β-carotene or plasma β-carotene to vitamin A status will remain clouded. Recent estimates of the vitamin A value of plant foods suggest that in some cases plant-based diets may be insufficient to improve vitamin A status, yet in some circumstances this prediction has proved inaccurate. This paper will contrast several of the more recent methods of assessing the bioavailability of food-borne provitamin A carotenoids to illustrate the potentially complementary nature of the two categories of bioavailability model.