TSRC, Tob. Sci. Res. Conf., 2017, 71, abstr. 078

Development of a multi-stage liquid extraction method for the quantitative analysis of flavors in tobacco filler

U.S. Food and Drug Administration, Atlanta, GA, USA

A method to quantify flavor chemicals in the filler of combustible tobacco products was validated using GC-MS. The tobacco filler samples from cigarettes, little cigars, cigarillos, and cigars were prepared in a multi-step extraction procedure using ethyl acetate and water followed by dilution with hexane. The features of the extraction stages include: extraction of non-polar compounds into ethyl acetate, water addition for polar compounds, partitioning into two-layers with subsequent centrifugation, followed by the removal of the ethyl acetate layer and mixing with an equal volume of hexane containing the internal standard (IS), second centrifugation, yielding an extract for GC-MS analysis. The selected flavor chemicals are quantitated by GC-MS in full scan mode by relating the peak area ratios to the concentration using target compound extracted ion chromatograms (EIC) compared to the IS peak area.

The linearity, accuracy, precision, limit of quantitation (LOQ), and selectivity results were used to demonstrate the extraction and instrumental analysis was appropriate for the quantitative analysis. The linearity of the method was assessed across the range of 10 to 300 ppm (μg/g). The accuracy and precision was evaluated using replicate measurements of matrix spiked samples at 50, 100, and 150 ppm levels, in addition to the LOQ standard at 10 ppm. The LOQ was established by evaluating the precision of all compounds, <10% relative standard deviation (%RSD), as well as the signal-to-noise ratio for the lowest absolute peak area EIC, S/N≥11. The accuracy ranged from 95-127% for individual replicates and the precision at each concentration level ranged from 3-8% RSD. The selectivity results, which required no interfering peaks in the blank matrix and sufficient resolution at the LOQ level, met the validation criteria by visual interpretation and illustrated for each EIC. In addition to the quantitative analysis of the selected flavor compounds, the use of the full scan mode can also provide identification and library searching capabilities for unknown compounds. The results from the validation support the use of the proposed method for quantitative analysis of flavor related compounds in unburned portion of combustible tobacco product filler.