5. AOACSPDSMethods-2018AwardsV3

58

K oshy et al .: J ournal of AOAC I nternational V ol . 99, N o . 6, 2016  1453

B

A

C

D

Figure 9. Linearity of response for withanolide B over the full concentration range by (A) linear regression and (B) residual analysis and over a truncated concentration range by (C) linear regression and (D) residual analysis.

error cannot be determined, though the sum of withanolides is generally higher under condition 2 compared with condition 1. The differences are small enough, however, that the method is considered rugged based on intermediate precision.

withanolide in each material varying from 0.15 to 1.10% w/w (1500–11000 ppm), the acceptable range of recovery is 95–105% (Table 2). Five of the six withanolide analytes were <95% recovery in at least one of the four materials, but none of the withanolide analytes was <95% recovery in all four materials. Recovery of total withanolide content frommaterials spiked after extraction (WS/06Lot08, WS/05Lot20, RD/1045, and ERH-46) varied from 91.0 to 99.6% for W-IV, W-V, WF-A, 12-D, and W-A. These results are very similar to the recoveries when the materials are spiked prior to extraction, demonstrating that the extraction process is efficient. The results for W-B were unexpectedly low in this experiment, at 70.7–72.2% recovery. A review of the experimental details did not uncover a likely reason for this low recovery, however it is noted that these values are not typical for this analyte. If the extraction of W-B were inefficient, one would expect lower recoveries for the materials spiked before extraction compared with those spiked after. The observed lower recoveries of W-B when materials were spiked after extraction suggests an error in the amount of W-B added, but this could not be confirmed.

Stability of Sample Solution

Stability data are presented in Table 14. There was no significant decrease in withanolide content for any of the analytes or the sum of withanolides in the sample extract over a 24 h period at 25±2°C.

System Suitability

System suitability parameters were determined throughout method development and tested during the validation study. The following suitability parameters define the optimal performance of the method: ( 1 ) Injection of mixed standard preparations in triplicate with every run; ( 2 ) peak resolution between withanoside V and withaferin A is >3; ( 3 ) the tailing factor for each analyte is <1.5; ( 4 ) the relative retention times compared with withanolide A are ~0.70 for withanoside IV, ~0.89 for withanoside V, ~0.92 for withaferin A, ~0.96 for 12-deoxywithastromonolide, and ~1.15 for withanolide B; ( 5 )–the repeatability precision is <2.0% RSD for standards and samples; ( 6 ) the retention time window is within 3σ of the mean retention time of the standard or QC-check sample; and ( 7 ) the QC-check sample result is within 10% of the label value. It is

Ruggedness

Table 13 shows the results of ruggedness testing. The intermediate precision on the sum of withanolides varied from 0.91 to 6.17%. Because the experiment was not performed as a Youden-type ruggedness trial, the primary source(s) of