Development of Kudzu (Pueraria Montana var. lobata) Reference Materials for the Determination of Isoflavones and Toxic Elements

Abstract

Background

In collaboration with the Office of Dietary Supplements at the National Institutes of Health, the National Institute of Standards and Technology issued a suite of botanical matrix reference materials (RMs) and Standard Reference Material^® (SRM) for determination of isoflavones and toxic elements in kudzu dietary supplement ingredients.

Objective

RM 8650 Pueraria montana var. lobata (Kudzu) Rhizome, SRM 3268 Pueraria montana var. lobata (Kudzu) Extract, and RM 8652 Kudzu-Containing Solid Oral Dosage Form were issued with values assigned for isoflavones (puerarin, daidzin, and daidzein), toxic elements (arsenic, cadmium, and lead), and selenium.

Methods

Isoflavone values were assigned using liquid chromatography with UV absorbance or mass spectrometry detection. Element values were assigned using inductively coupled plasma mass spectrometry and results from an interlaboratory comparison exercise.

Results

Mass fractions for puerarin were 32.2 ± 3.2 mg/g, 128 ± 13 mg/g, and 68.2 ± 6.9 mg/g in RM 8650, SRM 3268, and RM 8652, respectively. Arsenic increases from 156 ± 14 ng/g to 849 ± 83 ng/g and cadmium decreases from 348 ± 14 ng/g to 82.1 ± 4.9 ng/g from rhizome to extract.

Conclusion

The kudzu RM/SRM suite complements previously issued soy-related SRMs with values assigned for isoflavones, which have been studied for their potential health benefits, and expands the analytical resource by providing values for puerarin, an isoflavone not found in soy.

Highlights

The three new kudzurmaterials are for use in the determination of isoflavones, toxic elements, and selenium. For the isoflavones, these new kudzu materials provide higher levels of daidzin and daidzein than existing soy-related SRMs, and they provide a value for an isoflavone not in existing SRMs (puerarin). Toxic elements in RM 8650 and SRM 3268 provide new botanical matrixes for use by dietary supplement manufacturers for the verification of the safety of their raw materials.

Since 2002, the National Institute of Standards and Technology (NIST) has collaborated with the Office of Dietary Supplements at the National Institutes of Health to provide reference materials (RMs) for use in the validation of analytical methods for the determination of chemical markers and toxic elements in botanical dietary supplement ingredients (1). Standard Reference Materials^® (SRMs) and RMs have been developed for botanical dietary supplement ingredient matrixes including Ginkgo biloba (2), saw palmetto (3), green tea (4), St John’s wort, yerba mate tea, kelp (5), turmeric, and ginger as well as non-botanical supplements such as multivitamin tablets (6), fish and plant oils (7), iodized salt, calcium tablets, and chromium tablets. In 2014, NIST issued a suite of soy-related SRMs with values assigned for the content of isoflavones, which are plant-derived polyphenolic compounds that exhibit estrogenic or hormone-like activity. These soy-based SRMs included the following matrixes: soy flour, soy protein isolate, soy protein concentrate, soy-containing solid oral dosage form, and soy milk (8). These soy-matrix SRMs (with the exception of the soy milk) were issued with values assigned for six isoflavones including three aglygones (daidzein, genistein, and glycitein) and three glycosides (daidzin, genistin, and glycitin).

To complement the soy materials, NIST recently developed a suite of three kudzu-matrix RMs and an SRM including rhizome, rhizome extract, and kudzu-containing solid oral dosage form (SODF). Kudzu, also known as Japanese arrowroot (Pueraria montana var. lobata), is a perennial vine that is a member of the pea family and is native to eastern and southeastern Asia. Introduced into the United States in 1876 and heavily promoted to prevent soil erosion in the 1930s and 1940s, kudzu has a tradition of use as an herbal medicine, most notably for conditions related to alcohol consumption (9, 10). Kudzu contains three prominent isoflavones: daidzein, daidzin, and puerarin. Daidzein and daidzin are present in soy; however, puerarin, which is the most abundant isoflavone in kudzu, is not found in soy products. Kudzu preparations are marketed as dietary supplements in the United States (9, 10), and research indicates that puerarin may increase blood flow to the heart and brain (11). However, most studies on kudzu’s health effects were small in scale and had weak experimental designs (12–15); therefore, further research is necessary to draw valid conclusions regarding potential benefits of kudzu as a dietary supplement. An important consideration for future trials to evaluate potential effects of kudzu and isoflavone intake on human health is accurate quantification of study materials, for which reference materials are a critical tool.

From the standpoint of labeling and/or standardization of kudzu supplements, the isoflavones are the constituents of interest regarding perceived health benefits. However, from a safety aspect, a concern for many botanical dietary supplements is often the toxic element content, specifically arsenic, cadmium, lead, and mercury. Arsenic is ubiquitous throughout the environment in the soil, water, and air. Generally regarded as a toxic element, arsenic toxicity is dependent on its chemical form and oxidation state with inorganic arsenic being considered most toxic and the organic forms considered non-toxic or only mildly toxic (16). Cadmium and lead mainly enter the food chain from anthropogenic activities. For cadmium, these activities include processing of ores, burning of fossil fuels and municipal waste, and the application of phosphate and sewage sludge-containing fertilizers (17). Mercury is a toxic element that has been widely emitted into the environment from industrial activities and consumer products. A major route of exposure is from dietary intake, mostly in the form of methylmercury in seafood. Exposure can occur from sources such as animals fed with fish meal, contaminated crops, and dietary supplements, including fish oils (18).

RM 8650 Pueraria montana var. lobata (Kudzu) Rhizome, SRM 3268 Pueraria montana var. lobata (Kudzu) Extract, and RM 8652 Kudzu-Containing Solid Oral Dosage Form (SODF) were issued with values assigned for mass fractions of puerarin, daidzein, and daidzin. The three isoflavones were determined using an LC with UV absorbance detection and MS detection (LC–UV–MS) method, which was adapted from a method developed by Phillips et al. (8) to determine isoflavones in the soy-matrix SRMs. The kudzu-matrix RMs and SRM complement the soy-related SRMs by providing different matrixes, higher levels of daidzein and daidzin, and values for an additional isoflavone with putative health effects, puerarin.

SRM 3268 Pueraria montana var. lobata (Kudzu) Extract was issued with certified mass fraction values for arsenic, cadmium, lead, and a nutrient element, selenium, based on measurements using inductively coupled plasma mass spectrometry (ICP-MS) at NIST and results from collaborating laboratories. For RM 8650 Pueraria montana var. lobata (Kudzu) Rhizome, reference mass fraction values were assigned for arsenic, cadmium, and lead using ICP-MS. SRM 3268 (extract) and RM 8650 (rhizome) complement existing botanical dietary supplement SRMs and RMs for toxic element content, particularly for arsenic levels, and SRM 3268 provides information on content of the nutrient element selenium. Although the primary intended use of the kudzu RMs/SRM is for method validation and quality assurance of isoflavone and toxic element measurements, plant identity information is also provided based on a molecular approach comparing associated DNA sequence analysis from botanically authenticated Pueraria montana var. lobata.

Experimental

Production of the RMs and SRM

The kudzu plant material was harvested in July 2008 in Turtle Creek, Boone County, WV, USA (latitude 37.98465; longitude 81.95446; elevation 226 m) by a trained botanist accompanied by representatives of the growing area and from Naturex (South Hackensack, NJ). The botanist revisited the collection site at a later date, when the kudzu plant flowered, to confirm a positive identification. The kudzu plant material was shipped to Naturex, where the dried ground root material (10 kg) and extract (2.25 kg, excipient is maltodextrin) were prepared. The kudzu ground root powder and extract were shipped to High-Purity Standards (Charleston, SC) for blending and packaging. The root powder and extract were each blended for 30 min by a rocking and rolling technique. After blending the extract was passed through a No. 35 (500 µm) sieve. Samples of the root powder (3 g) and extract (1 g) were heat sealed in nitrogen-flushed, 7.6 cm × 12.7 cm, 4 mil polyethylene bags and then sealed in nitrogen-flushed aluminized plastic bags with two packets of silica gel. A total of 3300 packets of kudzu rhizome powder (400 packets in eight boxes and 100 in the ninth box) and 1963 packets of kudzu extract (four boxes) were packaged as RM 8650 and SRM 3268, respectively. Sales units of RM 8650 kudzu rhizome and SRM 3268 kudzu extract both contain 5 packets.

RM 8652 Kudzu-Containing SODF was prepared by blending four different brands of commercially available kudzu products for a total of 6.3 kg. Three of the products contained blends of kudzu root extract (standardized to 1% or 2% daidzin) and kudzu root; the fourth product, which represented 57% of the mass of the final mixture, contained only kudzu root extract. The details of the commercial kudzu products used in RM 8652 are summarized in online Supplemental Table S1. The products were blended for 1 h in a mixing vessel and then packaged (2.6 g samples) as described above resulting in 1600 packets in four boxes. The three kudzu RMs were ⁶⁰Co-irradiated by Neutron Products, Inc. (Dickerson, MD) to an absorbed dose of 6.8 kGy to 8.5 kGy for RM 8650 (rhizome) and SRM 3268 (extract) and 7.4 kGy to 9.0 kGy for RM 8652 (SODF).

For the analysis of SRM 3268 for both toxic elements and isoflavones, eight sample packets were selected from the four boxes and were identified by the box number and bottom (B), middle (M), and top (T), e.g., 4B = box 4 bottom, and the duplicate subsamples from each packet were analysed (i.e., a and b). For the analysis of RM 8650 and RM 8652, one sample packet was selected from each of eight boxes and analyzed in duplicate and the samples are denoted as 1a, 1b, 2a, 2b, etc., where the number denotes the box and the duplicate subsamples from the same sample packet are denoted as a and b.

Taxonomic Identification

The taxonomic identity of the Pueraria montana var. lobata collected for preparation of RM 8650 was established through identification by a trained botanist using an herbarium specimen from the original material and from a molecular approach by comparing associated DNA sequence analysis from botanically authenticated Pueraria montana var. lobata. The experimental approach for assignment of identity based on DNA sequence analysis is described in detail in the Supplemental Information and in Tables S2–S6.

Materials

1. Calibrants.—The isoflavone standards, isotopically labeled analogues, and other compounds used as internal standards (manufacturer’s stated purity in parentheses) were obtained from commercial sources as follows: puerarin (98%; Indofine, Hillsborough, NJ); daidzin (91.56%) and daidzein (98.78%; Blaze Science Industries, Lawndale, CA); ¹³C₆-daidzin (97.5%) and ¹³C₆-daidzein (98.0%; IsoSciences, King of Prussia, PA); and caffeine (purity not determined; Sigma, St. Louis, MO). The purity of the isoflavone reference standards was assessed at NIST for puerarin (98.70%) using LC and for daidzin (91.66%) and daidzein (98.15%) using quantitative magnetic resonance (qNMR). The NIST purity assessment results were combined with the manufacturers’ stated purity to provide an average purity for each compound, i.e., puerarin (98.35%), daidzin (91.61%), and daidzein (98.46%), which was used to correct the results for purity.  For elemental analysis, SRM 3103a Arsenic (As) Standard Solution, SRM 3108 Cadmium (Cd) Standard Solution, SRM 3128 Lead (Pb) Standard Solution, SRM 3133 Mercury (Hg) Standard Solution, SRM 3149 Selenium (Se) Standard Solution, SRM 3102a Antimony (Sb) Standard Solution, and SRM 3120a Germanium (Ge) Standard Solution were used as calibrants or as internal standards. All SRMs were obtained from the NIST Office of Reference Materials (Gaithersburg, MD).

2. Reagents.—Sodium hydroxide and acetic acid used in the hydrolysis were reagent grade. High-purity grade dimethylsulfoxide (Burdick & Jackson, Muskegon, MI) and ammonium acetate (Sigma) were used for calibrant and mobile phase preparation, respectively. All other solvents were HPLC grade. An methanol-water 80 + 20 (volume fraction) solution was prepared and used as the extraction solvent to remove the isoflavones.

3. Control materials.—SRM 3238 Soy-Containing Solid Oral Dosage Form (now RM 8168) was used as a control material during the analysis of the kudzu RMs and SRM for the determination of isoflavones. SRM 3254 Green Tea (Camellia sinensis) Leaves, SRM 1566a Oyster Tissue, or SRM 1573a Tomato Leaves were used as control materials for the analysis of arsenic, cadmium, lead, mercury, and selenium. All SRMs were obtained from the NIST Office of Reference Materials.

Determination of Isoflavones

1. Preparation of calibrant solutions.—One stock solution each for caffeine, ¹³C₆-daidzin, and ¹³C₆-daidzein was prepared. Four stock solutions of puerarin, daidzin, and daidzein were independently prepared. The isoflavone was first dissolved in dimethylsulfoxide (DMSO) [approximately 0.1% to 0.5% (mass fraction) isoflavone-DMSO] before adding extraction solvent. Portions of these stock calibration solutions were gravimetrically combined with the internal standard solutions in ratios to resemble the concentration of each isoflavone in the SRM following extraction to determine the response factors. All solutions were stored at −20°C when not in use. Response factors were determined for the three isoflavones using three or four working calibrant solutions as summarized in online Supplemental Table S7.

2. Sample preparation.—Sample preparation for the three kudzu matrixes was similar with only slight modifications for the extraction step. Each packet of the kudzu RM/SRM and the control material was thoroughly mixed prior to weighing a 10 mg sample (exact mass determined) into a 1mL polypropylene centrifugation tube. An appropriate volume of internal standard solution (approximately 0.50 mL, exact mass determined) and 3.5 mL extraction solvent were added to the tubes and the contents mixed well. The samples were placed in an ultrasonic bath (60 Hz power and 40 kHz frequency) for 30 min without heating. Following ultrasonic extraction, the samples were centrifuged with (RCF = 1000 x g at 3000 rpm or 314.2 rad/s), at room temperature, for 10 min and the supernatant removed. Based on extraction optimization studies for each matrix, RM 8650 (rhizome) and RM 8652 (SODF) were extracted by means of a second 30-min sonication with 10 mL fresh extraction solvent and 10 min centrifugation. For SRM 3268 (extract), three additional extractions were performed with a fresh portion (5 mL) of extraction solvent added followed by 30 min sonication and 10 min centrifugation. Following each extraction cycle, the supernatant was combined to produce approximately 20 mL sample extract after the two or four extraction cycles. A 1.5 mL aliquot of 2 mol/L sodium hydroxide was added to the extract to hydrolyze any isoflavones that might be present as malonyl- or acetyl-glycosides to the corresponding glycoside. Following sonication for 15 min without heating, 500 µL glacial acetic acid was added to each sample to neutralize the base. A portion of the sample extract was placed in an autosampler vial for analysis by LC–UV–MS. A solvent blank, a sample blank (sample with no internal standards added), and an internal standard blank (internal standard solution with no sample) were also included in each sample set for analysis.

3. LC–UV–MS method.—The separation of the three isoflavones of interest in the kudzu materials was based on the LC method developed by Phillips et al. (8) to separate the six predominant isoflavones in soy-related materials using an amide column in a reversed-phase mode. Samples were analyzed by using an LC–UV–MS system from Agilent Technologies (Palo Alto, CA). The 1100 Series LC was equipped with a variable wavelength detector in line with an SL Series MS with electrospray ionization (ESI) source in the positive ion mode. An amide-embedded reversed-phase LC column (Ascentis Express RP-Amide column, 150 mm x 4.6 mm id, 2.7 µm particles, Supelco, Bellefonte, PA) was used for the analysis. The mobile phase consisted of two solvents: (A) 10 mmol/L ammonium acetate in water (pH 4.7) and (B) acetonitrile. Gradient elution used a solvent composition ramping from 5% B to 10% B in 30 min, then to 43.3% B in 20 min, followed by a 5 min wash at 90% B and 5 min re-equilibration to the initial conditions of 5% B. The flow rate was 1.2 mL/min with a column temperature of 35°C. The autosampler tray was maintained at 10°C and a 5 µL injection volume was used for all samples. During the first 15 min of the chromatographic analysis, the variable wavelength detector was used to monitor the elution of caffeine at 274 nm and then changed to 251 nm to monitor absorbance of the isoflavones. The MS was operated in ESI (positive ion mode) with a nebulizer pressure of 380 kPa (55 psi), a drying gas temperature of 340°C and flow rate of 12 L/min, vaporizer temperature of 240°C, a capillary voltage of 3000 V, corona current of 4 µA, charging voltage 2000 V, and fragmentor voltage of 110 V.

    Quantification was performed using an isotope dilution (ID) MS approach for two of the isoflavones, daidzin and daidzein, and an internal standard approach for puerarin with caffeine as the internal standard. For the IDMS approach, selected ion monitoring (SIM) mode was used with the following mass-to-charge ratios (m/z): puerarin (m/z 417), daidzin (m/z 417), ¹³C₆-daidzin (m/z 423), daidzein (m/z 255), and ¹³C₆-daidzein (m/z 261). The mass fraction for puerarin was determined using both the UV absorbance detection based on the internal standard caffeine and MS detection based on ¹³C₆-daidzin as the internal standard. The mass fractions of daidzin and daidzein were determined based on using ¹³C₆-daidzin and ¹³C₆-daidzein as internal standards for daidzin and daidzein, respectively.

    To assess homogeneity for the isoflavones in each kudzu material, duplicate samples were prepared from each of 10 packets of each RM or SRM, which were selected by a stratified random sampling scheme. Samples were randomized before both the sample preparation and the chromatographic analysis. Batch homogeneity for each compound was investigated through comparison of the mass fraction determined versus packaging order.

Determination of Arsenic, Cadmium, Lead, and Selenium Using ICP-MS

1. Preparation of calibrant solutions.—SRM 3103a Arsenic (As) Standard Solution, SRM 3108 Cadmium (Cd) Standard Solution, SRM 3128 Lead (Pb) Standard Solution, and SRM 3149 Selenium (Se) Standard Solution were diluted gravimetrically to prepare calibration curves ranging in mass fractions from 0.8 µg/kg to 20 µg/kg to ensure linearity of the calibration curves at the expected mass fractions for each element in solution. A matrix-matched standard addition spike, used for compensation of non-spectral interferences, was prepared gravimetrically at mass fractions of 200 µg/kg for arsenic and cadmium, at a mass fraction of 400 µg/kg for lead, and at a mass fraction of 30 µg/kg for selenium. An internal standard (IS), used to reduce the effects of instrumental drift, was prepared from SRM 3102a Antimony (Sb) Standard Solution, at a mass fraction of 1000 µg/kg and used for the determination of arsenic, cadmium, and lead, and an IS was prepared from SRM 3120a Germanium (Ge) Standard Solution, at a mass fraction of 1000 µg/kg for use in the determination of selenium. All standard solutions were prepared using 1.5% HNO₃ (volume fraction, in water).

2. Sample preparation.—Two 0.4 g subsamples from each of eight packets of candidate SRM 3268 and two 0.5 g subsamples from each of six packets of RM 8650 were prepared for analysis. Four 0.5 g subsamples of SRM 3254 Green Tea (Camellia sinensis) Leaves were prepared as control samples and used during the analysis of arsenic, cadmium, and lead. For the determination of selenium, two 0.5 g subsamples from each of four packets of SRM 3268 and three 0.5 g subsamples of SRM 1566a Oyster Tissue (control sample) were prepared for analysis. The subsamples were placed in digestion vessels and 9 mL concentrated HNO₃ and 1 mL concentrated HF were added to each vessel along with 0.30 mL 1000 µg/kg Sb IS solution to the samples analyzed for arsenic, cadmium, and lead and 0.25 mL 1000 ng/g Ge solution to the samples analyzed for selenium. The samples were digested using a microwave sample preparation system (CEM MARSXpress, Matthews, NC) at 800 W (100% power), 30 min ramp time at 150°C (15 min hold) followed by 85% power for 25 min at 185°C (15 min hold). After microwave digestion, the solutions were transferred to Teflon beakers and the solutions were heated on a hot plate (180°C) until reduced to near dryness. Samples were then transferred to polyethylene bottles and diluted to 30 g using 1.5% HNO₃ (volume fraction, in water) for analysis by ICP-MS.

    Samples were analyzed using ICP-MS (Agilent 7500cs, Palo Alto, CA) equipped with a Peltier-cooled, inert sample introduction system using H₂ as a collision gas to minimize polyatomic interferences. Arsenic, cadmium, lead, and selenium were measured according to the parameters in Supplemental Table S8. The argon flow on the ICP-MS was set to 15 L/min, the auxiliary flow to 0.8 L/min, and the nebulizer flow to 1 L/min. The radiofrequency (RF) power was set to 1500 W.

    Analyte mass fractions were calculated by the method of standard additions to compensate for matrix effects. Two aliquots were taken from the sample solution prepared and 0.2 g of the matrix matched standard addition spike was added to one of the aliquots. The amount of arsenic, cadmium, and selenium added to the spiked solution was 6 µg/kg, giving an estimated total mass fraction of arsenic and cadmium in the spiked solution of 9 µg/kg. The amount of lead added to the spiked solution was 12 µg/kg giving an estimated total mass fraction of lead in the spiked solution of 20 µg/kg. Results were corrected for the mean blank values by subtracting the mean total micrograms found in the blank solutions from the total micrograms found in each individual sample solution.

    To assess homogeneity for arsenic, cadmium, and lead, duplicate samples of 0.5 mg were prepared from each of six packets of RM 8650 and duplicate samples of 0.4 mg from each of eight packets of SRM 3268, which were selected by a stratified random sampling scheme and covered the entire lot of samples packaged. For the determination of selenium, duplicate samples were prepared from each of four packets of SRM 3268, which were selected by a stratified random sampling scheme. Samples were randomized before both the sample preparation and the sample analysis. Screening analyses for the toxic metals in RM 8562 kudzu SODF indicated that the mass fractions were sufficiently low that the assignment of values for this material would be of limited benefit.

Determination of Mercury Using ID Cold-Vapor (CV)-ICP-MS

1. Preparation of calibrant solutions.—The working ²⁰¹Hg isotopic spike solution used for calibration was prepared by accurate gravimetric dilution of a master stock solution that was calibrated by reverse ID using SRM 3133 Mercury (Hg) Standard Solution. Two separate stock solutions were prepared by serial dilution. Two spike calibration mixtures were prepared from each of these solutions generating four spike mixtures, and these solutions were measured using CV-ICP-MS under the same conditions as the samples.

2. Sample preparation.—Six subsamples of candidate SRM 3268 of nominal mass 0.28 g were accurately weighed by difference into a microwave quartz vessel and spiked with a weighed aliquot of ²⁰¹Hg (approximately 10 µg/kg) followed by the addition of 6 g of high-purity nitric acid. Five subsamples of SRM 1573a Tomato Leaves were analyzed as a control sample. The samples were digested in an Anton Paar Multiwave 3000 system (Graz, Austria) using a program from 0 W to 1400 W over 10 min, followed by a hold at 1400 W for 15 min, and a cool down period at 0 W for 15 min. After cooling to room temperature, the contents of the quartz vessels were transferred to 50 mL polypropylene centrifuge tubes and diluted for analysis by CV-ICP-MS.

    Mercury measurements were performed using ID CV-ICP-MS as described by Long et al. (19). The mercury vapor was generated by pumping sample and freshly prepared tin (II) chloride reductant solution (10% mass fraction in 7% volume fraction HCl) into a gas–liquid separator. The resulting Hg⁰ vapor was swept by the carrier gas into the ICP-MS (Thermo X7, ThermoFisher Scientific, Bremen, Germany). In addition to the samples and controls, five procedural blanks were carried through the entire processing and measurement scheme. The overall mean of these blank measurements was used to correct the sample measurement data.

    For initial homogeneity assessment of mercury across the material lot, single samples were prepared from each of six packets of candidate SRM 3268, which were selected by a stratified random sampling scheme. Samples were randomized before both the sample preparation and the sample analysis.

Interlaboratory Comparison for Elements

SRM 3268 Kudzu Extract was included as an unknown sample in the NIST Health Assessment Quality Assurance Program (HAMQAP) Exercise 2, which was conducted in 2018 (20). Between 17 and 23 laboratories reported results for the determination of arsenic, cadmium, lead, mercury, and selenium using a variety of analytical techniques. Each participating laboratory was requested to analyze a subsample from each of three packets of kudzu extract and to report the three results. NIST computed the average value and SD of the three measurements. At NIST, a consensus mean and consensus SD were calculated from the results received.

Dry-Mass Determination

The dry-mass conversion factor for each of the kudzu RMs and SRM was determined based on the combination of results for mass loss after drying in a forced air oven at 80°C and after drying in a desiccator over magnesium perchlorate [Mg(ClO₄)₂]. Two sets of 12 randomly selected packages for each of the three candidate RMs/SRM were obtained and labeled according to the boxes from which they were selected. For the desiccator studies, the loss of mass was determined after 5, 7, 14, 21, 28, 35, and 42 days. For the oven drying studies, the loss of moisture and volatile components was determined after drying for 1 h to 3 h at 80°C.

Results and Discussion

Taxonomic Identification

The taxonomic identity of Pueraria montana var. lobata used in RM 8650 was established using an approach used previously for four botanical SRMs, Ginkgo biloba, green tea (Camellia sinensis), St. John’s wort (Hypericum perforatum L.), and saw palmetto (Serenoa repens; 21). The definitions of rules for confidence estimates of species identity are provided in Supplemental Tables S2 and S3, and the reference samples and sequences used in the validation study (inclusivity and exclusivity panels) for RM 8650 (rhizome) are summarized in Supplemental Table S4. The trnL-F and the ITS2 DNA aligned matrixes for Pueraria montana var. lobata and relatives are provided in Supplemental Tables S5 and S6. The uncertainty associated with each nucleotide in the sequence and therefore the uncertainty associated with the DNA sequence as an identifier of species, is expressed in an ordinal scale that represents the confidence in the belief in the assigned value (0 = Most Confident, 1 = Very Confident, 2 = Confident, and 3 = Ambiguous; 22). Because the metrology for identity has not been fully developed, this approach represents a pragmatic way forward by considering these DNA sequences as the source of “comparability of identity” for the kudzu, i.e., chloroplast DNA sequences from authenticated kudzu are used to establish inclusivity and chloroplast DNA sequences from close relatives are used to establish exclusivity. DNA sequences for the kudzu rhizome are available in companion FASTA-formatted files (23). Based on the DNA sequence analysis, the confidence estimate for the species identification of RM 8650 as Pueraria montana var. lobata is denoted as “Confident (2)” (seeSupplemental Table S2 and S3 for the criteria for confidence estimates). For previously issued SRMs, this approach provided confidence estimates of “Most Confident (0)” for the Ginkgo biloba (24), St. John’s wort (25), and saw palmetto (Serenoa repens; 21) and “Ambiguous (3)” for the green tea (26).

Determination of Isoflavones

1. Extraction.—Extraction of the isoflavones from the three kudzu matrixes was based on previous work with soy-based matrixes (8). The extraction time was optimized by determining the effect of 30 min sequential extractions using fresh extraction solvent for each extraction. This procedure was repeated for each of the three kudzu matrixes. For SRM 3268 (extract), three extraction cycles were required before exhaustive recovery was observed for puerarin (see online Supplemental Figure S1). For daidzin and daidzein, no increase in recovery was observed after the first extraction cycle for SRM 3268. Based on these findings, four 30 min extraction cycles were used for SRM 3268. Based on similar extraction studies (results shown in online Supplemental Figures S2 and S3), two 30 min extraction cycles were determined to be exhaustive for the RM 8650 (rhizome) and RM 8652 (SODF) matrixes.

2. LC–UV–MS analysis.—For the determination of the three isoflavones in kudzu (puerarin, daidzin, and daidzein), the goal was to develop an LC–MS method and use an isotope dilution approach for quantification of the isoflavones. While ¹³C-labeled analogues of daidzin and daidzein were commercially available, an isotopically labeled analogue of puerarin was not available. Therefore, an LC method with sequential UV absorbance and MS detection was developed to quantify the puerarin using UV absorbance detection and the daidzin and daidzein using MS detection during the same chromatographic run.

    In the LC–UV absorbance method of Phillips et al. (8), sissotrin, an isoflavone not found in soy materials, was used as the internal standard for quantification. However, sissotrin was found to be present in SRM 3268 (extract) at about 4%, and therefore it was not suitable for use as an internal standard with the kudzu materials. Several other isoflavones (e.g., glycitin, glycitein, genistin, genistein, and ononin) were investigated for potential use as an internal standard. However, all eluted in regions of the chromatogram where other minor peaks were present as potential interferences. Ultimately, caffeine, which elutes earlier than the three isoflavones of interest, was selected as a suitable internal standard. Because of the early elution of caffeine, the mobile phase gradient program used by Phillips et al. (8) was modified to start with lower eluent strength and increase in eluent strength over a longer period.

    The LC–UV absorbance chromatogram from the determination of the isoflavones in SRM 3268 Pueraria montana var. lobata (Kudzu) Extract is shown in Figure 1. The LC–MS analysis of the isoflavones in RM 8650 Pueraria Montana var. lobata (Kudzu) Rhizome is shown in Figure 2. Similar chromatograms from the LC–UV absorbance and LC–MS analysis of calibrants and the remaining kudzu materials are provided in the online Supplemental Figures S4–S8. Response factors were determined for the three isoflavones based on three or four working calibrant solutions as summarized in online Supplemental Table S8. Using response factors, the mass fractions (mg/g) of the three isoflavones were determined on an as-received-basis, and the results are summarized in Table 1. The results of the individual measurements are provided in online Supplemental Tables S9–S11 for RM 8650 (rhizome), SRM 3268 (extract), and RM 8652 (SODF), respectively. The mass fraction for puerarin was determined from the same sample analyses using both the UV absorbance detection based on the internal standard caffeine and MS detection based on ¹³C₆-daidzin as the internal standard. The mass fractions of daidzin and daidzein were determined based on using ¹³C₆-daidzin and ¹³C₆-daidzein as internal standards for daidzin and daidzein, respectively.

    As shown in online Supplemental Tables S9–S11, the SD of the results for the three isoflavones in the three matrices ranged from <1% (puerarin by UV) to 5.5% (puerarin by MS). As expected, the lowest RSDs were observed for the UV absorbance measurements at 0.88, 0.94, and 1.9% for the SODF, extract, and rhizome, respectively. Results for daidzin and daidzein based on MS using an isotope dilution quantification approach had RSDs ranging from 1.0 to 4.5%. Three samples of SRM 3238 Soy-Containing Solid Oral Dosage Form were run as a control material during the analysis of each kudzu RM and SRM, and the results of these analyses are summarized in online Supplemental Table S12. For the analysis of the kudzu RMs and SRM, the results from the analysis of SRM 3238 were within the bounds of the assigned certified (diadzein) or reference value (diadzin) indicating that the analyses were in control. The mass fractions for puerarin determined using UV absorbance with the caffeine internal standard were consistently lower by 6 to 10% than the mass fractions determined using MS detection with the ¹³C₆-daidzin internal standard. Because the labeled daidzin elutes significantly later than the puerarin, it is not an ideal internal standard for quantification due to differences in MS ionization for the puerarin and ¹³C₆-daidzin resulting from a greater proportion of coeluting matrix components early in the separation.

3. Homogeneity for isoflavones.—The homogeneity of the kudzu RMs/SRM relative to isoflavones was assessed by analyzing duplicate subsamples of 10 mg from 10 packages of each material. The results of the determination of the three isoflavones in RM 8650 (rhizome) are plotted versus packaging order in Figure 3. Similar plots for SRM 3268 (extract) and RM 8652 (SODF) are provided in the Supplemental Figures S9 and S10, respectively. Analysis of variance at the 5% significance level was performed, and no trends were observed to indicate inhomogeneity at the 10 mg level for any of the isoflavones measured in the three kudzu-matrix RMs/SRM. Plots of mass fraction as a function of sample preparation order or LC analysis order also indicated no trends (results not shown).

4. Value assignment for isoflavones.—At NIST, certified values for organic constituents and trace elements in natural matrix SRMs are typically assigned based on the agreement of two or more independent analytical methods (27–29). For the kudzu-matrix materials, only one method was available for determination of the isoflavones daidzin and daidzein. Even though results were obtained for puerarin from two sets of measurements, i.e., LC–UV and LC–MS, the results were assigned as reference values rather than certified values because the two sets of measurements were not considered to be sufficiently independent to be categorized as certified values. The results for isoflavones in Table 1 were converted to a dry-mass basis using the results of the moisture determination summarized in online Supplemental Table S13. The assigned reference values for mass fractions of puerarin, daidzin, and daidzein (mg/g dry-mass basis) in the kudzu RMs and SRM are summarized in Table 2.

    For the isoflavones, the stated uncertainty incorporated uncertainty components due to moisture correction, purity correction, sample measurement variability, and a separate Type B uncertainty. The moisture-correction uncertainty derives from the estimation of the dry-mass proportion of the materials. The sample measurement variability is estimated with the standard error of the relevant mean. A Type B relative uncertainty component of 5% was determined to be appropriate to account for possible inhomogeneities in the measurements that might occur with this class of analytes for this group of materials. These components were combined in quadrature, consistent with the ISO Guide (30). In this case, the combined uncertainty is very much dominated by the largest component, which the Type B component. In contrast, the uncertainty components due to moisture correction and purity are extremely small in comparison and thus contribute very little marginally to the ensuing combined uncertainty.

    Screening measurements of the moisture content when the kudzu materials were received at NIST in 2012 agreed with the extensive measurements used as part of the value assignment process in 2018 (seeSupplemental Table S13). Therefore, the materials appear to be stable with respect to the moisture content. While the stability of the isoflavones in the kudzu RMs/SRM was not specifically investigated, isoflavones in soy matrixes have been determined to be stable for over 7 years. The stability of the isoflavones will be monitored regularly.

    As described previously (8), NIST has issued four soy-related SRMs with values assigned for six isoflavones, daidzin, daidzein, genistin, genistein, glycitin, and glycitein. Three of these materials (SRM 3236 soy protein isolate, SRM 3237 soy protein concentrate, and SRM 3238) soy solid oral dosage form) were recently changed to RMs (RM 8166, RM 8167, and RM 8188, respectively). The mass fractions of daidzin and daidzein in the kudzu materials and soy SRMs are compared in Figure 4. The kudzu RMs and SRM have higher mass fractions of diadzin and daidzein than the soy SRMs by one to two orders of magnitude with the exception of daidzin in the soy flour, which is only a factor of 2 lower than the kudzu rhizome. Both the kudzu SODF and soy SODF materials have similar levels of daidzin. The kudzu RM/SRM suite thus complements the soy RM/SRM suite by providing materials with higher concentrations of several isoflavones; materials containing puerarin, which is not present in the soy materials; and additional matrixes (rhizome and extract). Together the soy and kudzu SRMs/RMs provide resources for manufacturers of products containing isoflavones to demonstrate accuracy in their label claims and for clinicians conducting research to establish links between isoflavone intake and human health outcomes.

Figure 1.

LC–UV absorbance analysis of an extract of SRM 3268 Pueraria montana var. lobata (Kudzu) Extract at 274 nm for the first 15 min to monitor caffeine and thereafter at 251 nm for the determination of isoflavones.

Figure 2.

LC–MS analysis of extract of RM 8650 Pueraria montana var. lobata (Kudzu) Rhizome for determination of isoflavones. (A) m/z 255 to monitor daidzein, (B) m/z 261 to monitor ¹³C₆-daidzein, (C) m/z 417 to monitor puerarin and daidzin, and (D) m/z 423 to monitor ¹³C₆-daidzin.

Table 1.

Results for the determination of isoflavones in kudzu RMs and SRM using LC with UV absorbance and MS detection

	Mass Fraction (mg/g as-received basis)a
SRM	Puerarin (UV)	Puerarin (MS)	Daidzin (MS)	Daidzein (MS)
RM 8650 Rhizome	30.30 (0.57)	32.33 (0.69)	3.96 (0.13)	3.93 (0.18)
SRM 3268 Extract	121.9 (1.1)	136.2 (3.6)	7.84 (0.22)	16.34 (0.15)
RM 8652 SODF	64.47 (0.57)	68.3 (3.8)	11.38 (0.47)	3.320 (0.059)

^aUncertainty in parentheses is SD; corrected for purity of calibrants.

Figure 3.

Mass Fractions (mg/g) for (A) puerarin, (B) daidzin, and (C) daidzein in RM 8650 as a function of packaging order (box number) with duplicate sample preparation from the same package to determine homogeneity. Solid lines represent the average mass fraction value and the dash lines represent plus and minus one SD of the measurements. The blue dots represent the first subsample preparation and the red dots represent the second subsample preparation.

Table 2.

Mass fraction reference values assigned for isoflavones in kudzu dietary supplement RMs and SRM

RM/SRM	Name	Puerarin	Daidzin	Daidzein
		mg/g (dry-mass basis)a
RM 8650	Ground Kudzu (Pueraria montana var. lobata) Rhizome	32.2 ± 3.2b,c	4.21 ± 0.43b,d	4.17 ± 0.43b,d
SRM 3268	Kudzu (Pueraria montana var. lobata) Extract	128 ± 13b,c	8.21 ± 0.83b,d	17.1 ± 1.7b,d
RM 8652	Kudzu-Containing Solid Oral Dosage Form	68.2 ± 6.8b,c	12.0 ± 1.2b,d	3.51 ± 0.35b,d

^aConversion from as-received basis to dry-mass basis using the proportion (0.9421 ± 0.0033) gram dry-mass per gram as-received mass (RM 8650), (0.9543 ± 0.0003) gram dry-mass per gram as-received mass (SRM 3268), and (0.9460 ± 0.0018) gram dry-mass per gram as-received mass (RM 8652).

^bEach reference mass fraction value is the mean result from a single NIST analysis method using LC–UV absorbance or LC–MS. Values are expressed as x ± U_95%(x), where x is the estimated value and U_95%(x) is the expanded uncertainty of the value. The uncertainty of the estimate incorporates a component for moisture correction and a Type B relative component of 5%. The method-specific value of the analyte lies within the interval x ± U_95%(x) with about a 95% confidence (30, 31). The measurands are the total mass fraction of each isoflavone listed in Table 2, on a dry-mass basis, as determined by the method indicated.

^cAssigned value based on results from analysis using LC–UV absorbance.

^dAssigned value based on results from analysis using LC–MS.

Figure 4.

Comparison of mass fractions of isoflavones in kudzu RMs/SRM and soy SRMs. The uncertainties associated with the assigned value from the Certificate of Analysis or Report of Investigation are presented as relative percent (number above each bar in graph). The following SRMs and RMs are compared: RM 8650 Pueraria Montana var. lobata (Kudzu) Rhizome, SRM 3268 Pueraria Montana var. lobata (Kudzu) Extract, SRM 3234 Soy Flour, RM 8166 Soy Protein Isolate, RM 8167 Soy Protein Concentrate, and RM 8168 Soy-Containing Solid Oral Dosage Form.

Determination of Elements

1. ICP-MS analysis.—RM 8650 (rhizome) and SRM 3268 (extract) were analyzed at NIST using ICP-MS to determine the content of arsenic, cadmium, and lead and the results are summarized in Table 3. Selenium was also determined in SRM 3268 (extract) only. For RM 8652 (SODF), screening analyses for arsenic, cadmium, and lead indicated that the levels were below detection limits; therefore, further analyses of this material were not performed. Results for individual subsamples are provided in online Supplemental Table S14 for the arsenic, cadmium, and lead and in online Supplemental Table S15 for the selenium. As shown in Supplemental Table S14 for RM 8650, the relative expanded uncertainties for the ICP-MS measurements were 4.0% for cadmium, 8.8% for arsenic, and 30% for lead. For SRM 3268, the relative expanded uncertainties of the ICP-MS measurements were 4.9% for selenium, 5.5% for arsenic, 6.1% for cadmium, and 13% for lead. Results of the analysis of SRM 3254 Green Tea (Camellia sinensis) Leaves and SRM 1566 b Oyster Tissue, used as control samples during the analysis of RM 8650 (rhizome) and SRM 3268 (extract) are summarized in online Supplemental Tables S16 and S17. The results of the analysis of SRM 3254 and SRM 1566 b for arsenic, cadmium, lead, and selenium were within the bounds of the assigned values indicating that the analyses were under control.

2. ID CV-ICP-MS.—Mercury was determined in SRM 3268 (extract) using ID CV-ICP-MS, and the results are summarized in Table 3. Individual sub-sample results for Hg are provided in online Supplemental Table S19. Control results for Hg in SRM 1573a Tomato Leaves, as summarized in online Supplemental Table S20, were consistent with the certified value indicating that the analyses were under control.

3. Homogeneity assessment for toxic elements.—Homogeneity of the kudzu rhizome RM and kudzu extract SRM was assessed for arsenic, cadmium, and lead, and the results are summarized in Supplemental Table S14 and shown in Supplemental Figures S11–S13 for RM 8650 (rhizome) and online Supplemental Figures S14–S16 for SRM 3268 (extract). For all three toxic elements, the rhizome appears to be more homogeneous than the extract. Based on the results for lead in two of the samples of SRM 3268 (seeSupplemental Table S14), a subset of eight samples was reanalyzed as shown in Supplemental Table S18 with a mean of 1.03 mg/kg (SD = 0.15 mg/kg) which is nearly identical to the 1.04 mg/kg (SD = 0.24 mg/kg) for the original set of duplicates of eight samples (seeSupplemental Table S14).

    The results for mercury in SRM 3268 showed marked inhomogeneity in the material. Visually, the material was a fine powder with no discernible particulate dispersion. Mercury can be inhomogeneous in materials depending on the speciation and particle size distribution. The mass fraction values for mercury in SRM 3268 ranged from 5.0 ng/g to 16.0 ng/g (seeSupplemental Table S19) with a mean value of (10.4 ± 5.0) ng/g (48% RSD, n = 6). Results for mercury in SRM 3268 obtained from solid sampling atomic absorption spectrometry (AAS) screening also indicated a lack of homogeneity for the kudzu extract with a mean value of (7.8 ± 2.4) ng/g (31% relative, n = 3), similar to the mass fraction value reported in Table 3 and Supplemental Table S19 (see online Supplemental Information for experimental details of Hg screening analysis).

4. Interlaboratory comparison.—SRM 3268 (extract) was included as an unknown sample in the NIST HAMQAP Exercise 2, which was conducted in 2018 (20). In addition to the toxic elements, NIST also requested results for the determination of the selenium content as part of an effort to include elements of nutritional interest with values assigned in botanical dietary supplement SRMs. Currently, only the green tea leaf and extract SRMs have values assigned for nutritional elements (i.e., Al, Cu, Fe, Mn, and Zn). From the interlaboratory exercise, between 20 and 24 laboratories reported results for the determination of arsenic, cadmium, lead, mercury, and selenium using either ICP-MS or ICP-OES as their analytical technique. Seventeen laboratories reported results for the determination of mercury using either ICP-OES, ID-CV-ICP-MS, or cold vapor atomic absorption spectroscopy (CV-AAS) as the analytical technique. Laboratories also used a variety of sample preparation techniques including microwave and hot block digestions (the most common techniques used), acid hydrolysis, and open beaker digestion. Results from the interlaboratory comparison were compared to results from NIST determinations as shown in Table 3. For all elements, the consensus values determined using the participating laboratories’ data were within the NIST range of tolerance, which encompassed the target value bounded by twice its uncertainty.

5. Value assignment for elements.—The assigned certified and reference values for mass fractions of arsenic, cadmium, lead, and selenium (mg/g dry-mass basis) in SRM 3268 and RM 8650 are summarized in Table 3. Certified values for arsenic, cadmium, lead, and selenium were assigned in SRM 3268 based on the combination of the mean results of NIST analyses using ICP-MS and the weighted median of the individual laboratory means from the interlaboratory study. The combined value is the mean of the two method estimates (where available). The uncertainty of the combined mean incorporates both the within-method and between-method uncertainties, as well as a (very small) component related to moisture correction, and is estimated using a Monte Carlo procedure consistent with Supplement 1 of the ISO Guide (31, 32). The uncertainty of the mean of the ICP-MS results is the standard error of that mean. Due to the marked differences in results often present in interlaboratory exercises, the result for each element is the weighted median of the individual laboratory means for that element, where the weights are based on a Laplace random effects model (33). For this data, the weighted median is equal to or very close to the unweighted median of laboratory means for all elements. The uncertainty of the weighted median is estimated using a Monte Carlo procedure based on a Laplace random effects model for the between-lab and within-lab effects (31–34). Even though results for mercury in SRM 3268 were available from NIST using ID CV-ICP-MS and the interlaboratory comparison exercise, a value was not assigned, due to inhomogeneity.

    For RM 8650 (rhizome), reference values were assigned for arsenic, cadmium, and lead based on the mean results of NIST analyses using ICP-MS (seeTable 3). The reference values are expressed as x ± U_95%(x), where x is the estimated value and U_95%(x) is the expanded uncertainty of the value. The method-specific value of the analyte lies within the interval x ± U_95%(x) with about a 95% confidence (30, 31). Stability of the elements reported in the kudzu RMs/SRM has not been specifically investigated; however, based on experience with similar botanical SRMs, the materials are expected to be stable. The stability will be monitored regularly as these kudzu RMs/SRM are used as controls in other analyses or as unknown samples in quality assurance exercises such as HAMQAP.

    Of particular interest for the toxic elements is the relationship of the results in the rhizome (RM 8650) compared to the extract (SRM 3268), because the extract was prepared from the same collection of rhizome material. With results for the toxic element content in both materials, the question of whether the extraction process enriches the toxic element content in the extract or whether the elements are not efficiently extracted from the rhizome powder can be addressed. Comparison of the results of toxic elements in the rhizome with those in the extract indicates that the arsenic mass fraction is enriched by approximately a factor of 5 in the extract, whereas the cadmium mass fraction decreases by approximately a similar factor in the extract; the lead mass fraction remains at the same level in the extract as was found in the rhizome.

6. Comparison of toxic elements in kudzu SRM/RM to other botanical dietary supplement SRMs.—Several botanical dietary supplement SRMs and an RM exist with values assigned for toxic elements including Ginkgo biloba leaves (SRM 3246) and extract (SRM 3247; reference 2), green tea leaves (SRM 3254) and extract (SRM 3255; reference 4), St John’s Wort aerial parts (SRM 3262), kelp powder (SRM 3232; reference 5), ground turmeric rhizome (SRM 3299), and ginger rhizome (SRM 3398) and extract (RM 8666). The levels of toxic elements in the kudzu rhizome RM and extract SRM are compared to the existing botanical dietary supplement SRMs and RMs in Figure 5. The levels of lead in the two kudzu materials are similar to the lead content in the Ginkgo biloba leaves, St. John’s Wort aerial parts, kelp, turmeric rhizome, and ginger rhizome. Likewise, the level of arsenic in the kudzu rhizome is similar to the levels in green tea leaves, green tea extract, and the St. John’s wort aerial parts; however, the kudzu extract has the third highest mass fraction of arsenic (849 ng/g ± 83 ng/g) in any of the botanical dietary supplement SRMs available. The two SRMs with the highest levels of arsenic, kelp powder and ginger rhizome, have exceptionally high levels that are factors of about 50 times (38 300 ng/g ± 1300 ng/g and 49 600 ng/g ± 2600 ng/g) the mass fraction in the kudzu extract. The cadmium level in the kudzu rhizome is similar to the levels in the kelp powder and the St. John’s wort, which is known to accumulate cadmium. The mass fraction of mercury found in the kudzu extract (10.4 ng/g ± 5.0 ng/g) is similar to the ginger extract (RM 8666) (8.3 ng/g ± 2.6 ng/g). Beside selenium in the SRM 3268, NIST has provided reference values for the nutrient elements, aluminum, copper, iron, manganese, and zinc in SRM 3254 Green Tea (Camellia sinensis) Leaves and SRM 3255 Green Tea (Camellia sinensis) Extract. These additional elements were included in the SRMs because of interest to food manufacturing communities.

Table 3.

Mass fractions of elements in kudzu extract (SRM 3268) and kudzu rhizome (RM 8650)

	As	Cd	Pb	Hg	Se
	SRM 3268 Kudzu Extract—Mass Fraction (ng/g)
ICP-MS (n = 16: as-received basis)	771 ± 47a	81 ± 4a	1040 ± 140a, b		117 ± 5a
ID CV-ICP-MS (n = 6; as-received basis)				10.4 ± 5.0a
Interlab median (as-received basis)	849 ± 130c	74.7 ± 13c	1129 ± 320c	9.0 ± 6.2c	185 ± 110c
Certified value (dry-mass basis)d	849 ± 83e	82.1 ± 4.9e	1163 ± 155e		158 ± 73e
	RM 8650 Kudzu Rhizome—Mass Fraction (ng/g)
ICP-MS (n = 10; as-received)	147 ± 18a	328 ± 18a	1092 ± 55a
Reference value (dry-mass basis)d	156 ± 14f	348 ± 14f	1159 ± 46f

^aExpanded uncertainty at approximately 95% level of confidence.

^bResult of 1030 ± 150 (SD), as-received basis, was obtained for a second set of eight subsamples.

^cThe weighted median of each individual laboratory’s mean is used and the uncertainty is estimated using a bootstrap procedure (32). Both are based on a Laplace random effects model for the between-lab and within-lab effects (33).

^dConversion from as-received basis to dry-mass basis using the proportion (0.9543 ± 0.0003) g dry-mass/g as-received mass for SRM 3268, and (0.9421 ± 0.0033) g dry-mass/g as-received mass for RM 8650.

^eEach certified mass fraction value is the mean result of NIST analyses using ICP-MS and the weighted median of each individual laboratory’s mean. Values are expressed as x ± U_95%(x), where x is the certified value and U_95%(x) is the expanded uncertainty of the certified value. The true value of the analyte lies within the interval x ± U_95%(x) with 95% confidence. The measurands in Table 3 are total mass fractions for each analyte reported and metrological traceability is to the International System of Units (SI) derived unit for chemical mass fraction expressed as ng/g.

^fEach reference mass fraction value is the mean result of NIST analyses using ICP-MS. Values are expressed as x ± U_95%(x), where x is the estimated value and U_95%(x) is the expanded uncertainty of the value. The method-specific value of the analyte lies within the interval x ± U_95%(x) with about a 95% confidence (30, 31). The measurands are the total mass fraction of each element listed in Table 3 on a dry-mass basis as determined by the method indicated.

Figure 5.

Comparison of mass fraction (ng/g) of arsenic, cadmium, lead, and mercury in kudzu rhizome and extract RMs with available botanical dietary supplement ingredient SRMs (note that the y-scale is split in two locations). The uncertainties associated with the assigned value from the Certificate of Analysis or Report of Investigation are presented as relative percent (number above each bar in graph). The following SRMs and RMs are compared: SRM 3246 Ginkgo biloba Leaves, SRM 3247 Ginkgo biloba Extract, SRM 3254 Green Tea (Camellia sinensis) Leaves, SRM 3255 Green Tea (Camellia sinensis) Extract, SM 3262 St. John’s Wort (Hypericum perforatum) Arial Parts, SRM 3299 Turmeric (Curcuma longa L.) Rhizome, SRM 3398 Ginger (Zingiber officinale) Rhizome, RM 8666 Ginger (Zingiber officinale) Extract, SRM 3232 Kelp Powder, RM 8650 Pueraria Montana var. lobata (Kudzu) Rhizome, and SRM 3268 Pueraria Montana var. lobata (Kudzu) Extract.

Conclusions

Three new kudzu materials (two RMs and an SRM) have been issued by NIST for use in the determination of isoflavones, toxic elements, and selenium. For the isoflavones, these new kudzu materials provide higher levels of daidzin and daidzein than the existing soy-related SRMs, and they provide a value for an isoflavone not present in the soy materials (i.e., puerarin). For the toxic elements, both RM 8650 (rhizome) and SRM 3268 provide new botanical matrixes for use by dietary supplement manufacturers who need to verify the safety of their raw materials and fill a gap in existing SRMs/RMs regarding mass fractions of arsenic and the nutrient element selenium.