Evaluation of the probabilistic distribution of dietary biotin intake in Japan using Monte Carlo simulation.

Biotin is a widely distributed water soluble vitamin. Adequate intake of biotin was set at 50 µg/d in Japan 2010. Recently, the importance of the application of probabilistic techniques to estimate the share of the population at risk of deficient and excessive nutrient intake has been increasingly emphasized for assessing nutrient adequacy. Monte Carlo simulation, a computer-based method of analysis that uses statistical sampling techniques yielding a probabilistic approximation to the solution of a mathematical model, has been used to estimate the probabilistic distribution of the dietary intake of food chemicals. For this study, we used two preliminary models to estimate the dietary biotin intake with food consumption data based on the National Health and Nutrition Survey in Japan. One is evaluated by biotin concentration data from the total diet study; the other is a dataset of biotin concentration in individual foods. After removing outliers from the individual foods dataset, probability density distributions from two models showed analogous mean, median, 5th percentile, and 95th percentile values. The daily biotin intakes from these probabilistic methods showed that more than 80% of the Japanese population had higher than the adequate intake of biotin. However, the contribution of each food group to the total daily biotin intake was somewhat different. Improvement of these methods necessitates the collection of more actual data associated with sample compositional variability and evaluation of uncertainty associated with the food group classification of biotin.