5. AOACSPDSMethods-2018AwardsV3

120

14  B rown & Y u : J ournal of AOAC I nternational V ol . 96, N o . 1, 2013

photosensitivity, and extraction solvent volume. Variables that had significant effects on the total ginsenosides were sample weight, percent organic component in the extraction solvent, and mixing.

Table 2. HPLC gradient elution timetable for the determination of ginsenosides

MPA, % a

MPB, % a

Time, min

0

76

24

Collaborative Study Design

8

76

24

18

68

32

The SLV method was provided to 14 laboratories participating in the interlaboratory study. A total of 12 interlaboratory test samples were provided to each laboratory. Random identification numbers were assigned to each of the materials. Five of the test materials were submitted as blind duplicates: two test materials were prepared using a P. ginseng whole root powder and a P. ginseng powdered extract provided by the Canadian Phytopharmaceutical Corp., Richmond, BC, Canada; two were prepared using a P. quinquefolius ginseng whole root and powdered extract sample provided by Anthony Windust from the National Research Council of Canada (NRC) Ottawa, ON, Canada; and one was treated as a negative control consisting of maltodextran (99%) and magnesium stearate (1%) spiked with P. ginseng powdered extract. This negative control was typical of ginseng capsule fillers on the U.S. market. The remaining two test materials, a P. ginseng powdered extract in hard-filled capsule and a P. quinquefolius whole root sample in tablet form, were finished product dosage forms containing ginseng extract or powdered botanical, alone or in combination with other dietary supplement ingredients. These finished products are unique in appearance and as such were submitted as single samples not blind duplicates. All test samples were blinded in terms of content and identity. Each laboratory was provided with two practice samples: one P. quinquefolius powdered root sample and one P. ginseng powdered extract. The laboratories were required to complete the practice samples prior to analysis of the 12 test samples in order to become familiar with the extraction method and analysis. A mixed ginsenoside reference standard (MGRS) consisting of six ginsenosides in methanol solution (Rg 1 , Re, Rb 1 , Rc, Rb 2 , and Rd) at varying concentrations was also supplied to each laboratory. Each laboratory prepared its working standards from this MGRS according to the instructions included in the prescribed method. A total of 14 laboratories from three countries participated in the study. The laboratories cooperating in the study were recruited based on availability of equipment, expertise in ginseng or other herbal analysis, and an interest in the topic. All practice sample and system suitability results were evaluated to ensure that each laboratory was thoroughly familiar with the method of analysis. Laboratories that encountered technical difficulties with the method were required to reconduct the analysis of the practice samples. Participating Laboratories

25

60

40

42

52

48

43

0

100

44

0

100

45

76

24

a  MP = Mobile phase.

multiple replicates of each test sample. Four replicates of each matrix material were prepared and analyzed on 3 separate days, for a total of 12 replicates for each material. Within-day, between-day, and total SDs were calculated for each individual ginsenoside for each of the 12 replicates using a single factor analysis of variance (ANOVA) with an α-value of 0.5. The Horwitz Ratio (HorRat) value for each analyte in each material was also calculated. The ANOVA indicated that there was no significant difference for between-day precision. HorRat values fell within the acceptable range of 0.3–1.3 (18, 19). The accuracy of the method was demonstrated through two methods: re-extracting the residue of the sample used from the repeatability experiments and spiking a negative control that consisted of maltodextran (99%) and magnesium stearate (1%), with a well-characterized P. ginseng extract at three different levels (15, 40, and 75% extract by weight) over 3 days. All re- extracted replicates showed no detectable levels of any of the analytes present. The average spike recovery for the three levels was 101.5% with an RSD r of 3.24%. Accuracy

Youden Ruggedness Trial

The ruggedness of the method was assessed with a Youden ruggedness trial (20). The seven factors examined in this trial were sonication time, sample weight, percent organic component in the extraction solvent, mixing, extraction temperature,

Table 3. Concentration of each ginsenoside for each linearity (Lin) solution in the seven-point calibration curve

Concentration, µ g/mL Lin 1 Lin 2 Lin 3 Lin 4 Lin 5 Lin 6 Lin 7

Analyte

Test Sample Preparation

Rg1

394

131 77

45

15

9

5

Re

1188 396 231 135 46 27 16

Root samples were ground to a fine powder (approximately 50 mesh size) and homogenized prior to shipment. Powdered extracts were tested as-is. For tablet and hard-shell capsule finished product dosage forms, a minimum of 20 dosage units were combined and powdered by the participating laboratories to reduce variation due to sample inhomogeneity. The negative

Rb1

2981 994 580 338 115 67 39

Rc

499

166 97

57

19 11 7

Rb2

406

135 79

46

16

9

5

Rd

600

200 117

68

23 14 8