Carotenoid bioavailability from plant and animal food is highly variable depending on numerous factors such as the physical deposition form of carotenoids. As the carotenoid zeaxanthin is believed to play an important role in eye and brain health, we sought to compare the human bioavailability of an H-aggregated with that of a J-aggregated deposition form of zeaxanthin encapsulated into identical formulation matrices. A randomised two-way cross-over study with sixteen participants was designed to compare the post-prandial bioavailability of an H-aggregated zeaxanthin and a J-aggregated zeaxanthin dipalmitate formulation, both delivering 10 mg of free zeaxanthin. Carotenoid levels in TAG-rich lipoprotein fractions were analysed over 9·5 h after test meal consumption. Bioavailability from the J-aggregated formulation (AUC=55·9 nmol h/l) was 23 % higher than from the H-aggregated one (AUC=45·5 nmol h/l), although being only marginally significant (P=0·064). Furthermore, the same formulations were subjected to an internationally recognised in vitro digestion protocol to reveal potential strengths and weaknesses of simulated digestions. In agreement with our human study, liberation of zeaxanthin from the J-aggregated formulation into the simulated duodenal fluids was superior to that from the H-aggregated form. However, micellization rate (bioaccessibility) of the J-aggregated zeaxanthin dipalmitate was lower than that of the H-aggregated zeaxanthin, being contradictory to our in vivo results. An insufficient ester cleavage during simulated digestion was suggested to be the root cause for these observations. In brief, combining our in vitro and in vivo observations, the effect of the different aggregation forms on human bioavailability was lower than expected.

Production and manufacturing methods and the food carrier may influence the properties of probiotic strains, and have an impact on the outcome of clinical intervention studies. The aim of the present study was to establish whether the properties of a specific probiotic strain, Lactobacillus rhamnosus GG, may differ depending on the product and source of the strain. In total, fifteen different L. rhamnosus isolates, among them fourteen labelled as L. rhamnosus GG, were isolated from specific probiotic products. The micro-organisms were phenotypically and genotypically characterised. Their adhesion properties were compared using the human intestinal mucus model, and the ability of the isolates to influence model pathogen adhesion to human colonic mucus was assessed. All L. rhamnosus isolates used were confirmed as members of the species L. rhamnosus. Except the reference strain OL, all L. rhamnosus isolates showed randomly amplified polymorphic DNA, enterobacterial repetitive intergenic consensus and pulsed-field gel electrophoresis profiles identical to that of L. rhamnosus GG (ATCC 53103). All L. rhamnosus isolates showed similar tolerance to acid and were able to bind to human colonic mucus. However, pathogen exclusion by inhibition and competition varied significantly among the different L. rhamnosus isolates and pathogens tested. The results suggest that different sources of the same probiotic may have significantly altered strain properties. This should be considered in in vivo studies on human subjects and also for quality control of probiotic products.

Data on the vitamin A equivalency of β-carotene in food are inconsistent. We quantified the vitamin A equivalency (μg) of β-carotene in two diets using the dual-isotope dilution technique and the oral–faecal balance technique. A diet-controlled, cross-over intervention study was conducted in twenty-four healthy adults. Each subject followed two diets for 3 weeks each: a diet containing vegetables low in β-carotene with supplemental β-carotene in salad dressing oil (‘oil diet’) and a diet containing vegetables and fruits high in β-carotene (‘mixed diet’). During all 6 weeks, each subject daily consumed a mean of 55 (sd 0·5) μg [13C10]β-carotene and 55 (sd 0·5) μg [13C10]retinyl palmitate in oil capsules. The vitamin A equivalency of β-carotene was calculated as the dose-corrected ratio of [13C5]retinol to [13C10]retinol in serum and from apparent absorption by oral–faecal balance. Isotopic data quantified a vitamin A equivalency of [13C10]β-carotene in oil of 3·4 μg (95 % CI 2·8, 3·9), thus the bio-efficacy of the β-carotene in oil was 28 % in the presence of both diets. However, data from oral–faecal balance estimated vitamin A equivalency as 6:1 μg (95 % CI 4, 7) for β-carotene in the ‘oil diet’. β-Carotene in the ‘oil diet’ had 2·9-fold higher vitamin A equivalency than β-carotene in the ‘mixed diet’. In conclusion, this extrinsic labelling technique cannot measure effects of mixed vegetables and fruits matrices, but can measure precisely the vitamin A equivalency of the β-carotene in oil capsules.