6. AOACSPIFANMethods-2018Awards

224

C ampos G iménez & M artin : J ournal of AOAC I nternational V ol . 100, N o . 1, 2017  141

(c)  System suitability test .—Equilibrate the chromatographic system for ≥0.5 h. Inject a working standard solution of ascorbic acid at least six times and check the peak retention times and response (peak height or area). Ensure orotic acid and isoascorbic acid are fully resolved from ascorbic acid by injecting separate standard solutions of each compounds (prepared as stated for ascorbic acid). If the acids are not resolved, decrease the pH of the mobile phase to 5.0 or increase the amount of acetonitrile used. (d)  Calibration .—Perform single injections of working standard solutions, as a minimum at the beginning and end of each analytical series. Establish the calibration curve (seven points) by plotting peak response (height or area) vs ascorbic acid concentration, perform linear regression, and calculate the slope and intercept of the calibration curve. (e)  Analysis .—Perform single injections of sample solutions. (f)  Identification .—Identify the ascorbic acid peak in the chromatograms of the sample solutions by comparing it with the retention time and UV spectrum of the corresponding peak in the standard solution.

duplicates were assigned for analysis on the same day. Results were communicated to the Study Director using an electronic template similar to the one used in part 1. Part 3.— Two of the samples [infant formula RTF (milk-based) and adult nutritional powder (low-fat)] failed to meet acceptance criteria during part 2. Due to the questionable integrity of the samples (some laboratories reported product spoilage) and the fact that the samples had reached their expiration date, new products representing the same type of matrix were sent to a subset of 12 laboratories for analysis following the same protocol. Results were transmitted to the Study Director using the same electronic template as in parts 1 and 2. After data collection, outliers were detected using Cochran’s and Grubbs’ tests. Average concentrations, S r , and RSD r were estimated from the blind duplicates. S R , RSD R , and Horwitz ratio (HorRat) values (i.e., RSD R /predicted RSD R ) were also estimated. Details on statistical analysis can be found in Official Methods of Analysis SM “ Appendix D : Guidelines for Collaborative Study Procedures To Validate Characteristics of a Method of Analysis ” (7). Twenty-six laboratories initially agreed to participate in the collaborative study. Two laboratories dropped out during part 1 as a result of issues related to the availability of resources. The remaining 24 laboratories set up the method as described in the protocol. Fourteen laboratories used the UHPLC conditions as described in the protocol, whereas the remaining 10 used previously published HPLC conditions (5). No differences could be found between the provided results in either condition, and thus evaluation was performed combining all results. During method setup, it was brought to the attention of the Study Director that there was a need to establish suitability testing to ensure proper chromatographic separation between ascorbic, isoascorbic, and orotic acids. This suitability testing was added to the method as presented in this paper. One laboratory did not receive the practice samples due to customs restrictions. The laboratory qualified for part 2 by using results from the reference samples.After data compilation, average concentrations and S r and S R were calculated. Another laboratory reported single results and was not included in the statistical evaluation, although it qualified for part 2. Two laboratories reporting data above or below the average (±2×SD) were flagged as possible outliers and informed accordingly. Nevertheless, the two laboratories were accepted to continue because their results, although questionable, were still within ±3×SD. J. Statistical Evaluation Results and Discussion Part 1

H. Calculations

Calculate the concentration of vitamin C in mg ascorbic acid/100 g expressed in as-is ready-to-feed (RTF) products—or as reconstituted powder for powder samples—as follows:

(

)

A I V V S m V 1 3 2

− × × × × × ×

100

C

=

1000

where A is the response (height or area) of the ascorbic acid peak obtained for the sample solution, I is the intercept of the calibration curve, S is the slope of the calibration curve, m is the weight of the test portion in grams (2.0 g), V 1 is the volume of the test solution (volume used to dissolve the test portion) in milliliters (10 mL), V 2 is the volume used in the sample dilution (1.0 mL), and V 3 is the volume of the final sample dilution (10 mL). Note : If results expressed in the powder sample are needed, use the reconstitution rate for the calculation C× (225/25). Part 1.— All participant laboratories received two practice samples and were asked to analyze each of them in duplicate (two extractions from each reconstituted sample). Any deviation from the written method was to be recorded and reported. Results were communicated to the Study Director using the electronic template provided with the protocol. The participants were asked to report final vitamin C results, peak responses for standard curves and samples, and the different masses used during sample preparation. After review by the Study Director, results within a range of expected levels (average ±2× S R ) were used to identify the laboratories that qualified for part 2 of the study. Part 2.— All qualified laboratories received a second shipment containing 20 coded products, corresponding to 10 products in blind duplicates. The samples were a set of infant formula and adult nutritional products, representing a wide range of commercially available products. The laboratories were asked to analyze all the samples (a single extraction from each liquid or reconstituted powder) on 2 days (10 samples/day). Each sample was assigned to either days 1 or 2. The blind I. Collaborative Study Protocol

Part 2

Of the 24 laboratories providing results for practice samples and qualified to continue to part 2, two did not receive the full collaborative study set due to customs restrictions. The