Challenges in Developing Analytically Validated Laboratory-Derived Dietary Supplement Databases.

ABSTRACT

The Dietary Supplement Label Database (DSLD) is sponsored by the Office of Dietary Supplements (ODS) and the National Library of Medicine (NLM). It provides a searchable, free database of the contents of ∼65,000 supplement labels. A companion database of analytically verified product labels [the Dietary Supplement Ingredient Database (DSID)] was created by ODS, NLM, and the USDA. There are considerable challenges to populating both databases, but the DSID faces unique analytic chemistry challenges. This article describes the challenges to creating analytically verified marketplace surveys of dietary supplement (DS) product content claims for inclusion in public databases. Nutritionists and public health scientists require information on actual exposures to DS constituents because labeled content may not match labeled product content. Analytic verification of composition of DSs provides a link to actual exposure. A public database of analytically derived DS content was developed to provide more accurate estimates of dietary intake in population-based epidemiologic studies. The DSID has conducted surveys of several types of vitamin- and mineral-containing DSs. Results showing label content claims as analytically derived values are available in the current DSID. A recent pilot project explored the feasibility of adding botanical DS products to the DSID. Candidates for future botanical DSID studies will be based on sales volume, potential public health impacts, and the availability of validated analytic methods and reference materials. Databases like DSID and the DSLD are essential for researchers and clinicians to evaluate dietary ingredient intakes in population-based epidemiologic studies. Together, these databases provide a picture of the DS marketplace. The DSID provides an analytic survey of marketed DSs. However, selection of future botanical supplements for DSID evaluation involves analytic challenges. Even when appropriate resources are available, method selection and data evaluation are resource- and time-consuming.