6. AOACSPIFANMethods-2018Awards

275

D RAHER & W HITE : J OURNAL OF AOAC I NTERNATIONAL V OL . 101, N O . nnn , 2017 5

Figure 2017.03C. Typical sample chromatogram (NIST Standard Reference Material 1849a).

L. Instrumental Analysis and System Suitability

( 2 ) Residual error of linear regression. — The residual error for any one point of the four-point linear regression must not exceed 5%. In the event of excessive error, repeat the analysis with fresh standard solutions.

(a) Set the HPLC pump to a 0.5 mL/min flow rate and equilibrate the analytical column with mobile phase for 30 min. Set the column oven to 30°C. The instrument method designates an l ex of 295 nm and l em of 345 nm. The Agilent G1321A fluorescence detector lamp is designed to be on during analysis only due to the near-instant warm-up of the lamp. (b) Load standards, enzyme blanks, and samples onto the autosampler tray. The sequence includes a minimum of four initial injections for equilibration. The first of the four injections is a high standard to check that the Trp peak response does not exceed approximately 50% of full scale. (If significantly over 50%, lower the PMT gain by 1 unit. This will decrease the signal by about one-half. Results are best with a high standard between 20 and 50% of full scale.) The remaining three precision injections are 5.0/50 medium standard Figure 2017.03A and provide a check of system equilibration by calculating the percent difference in the last two middle standard injection Trp peak areas. The system is considered equilibrated if the areas agree within approximately 1%. If not equilibrated, an additional middle standard injection should be added and the system rechecked for area agreement. This is a prerun suitability check only. (c) Once the system is equilibrated, order the sequence with the four standard levels (high to low is recommended), blanks Figure 2017.03B, and samples, and then the bracketing standard injections. Repeat the calibration after approximately 20 sample injections. Following the sample analysis, rinse the analytical column with a solution of approximately 25% methanol in water. For longer storage, the column is best left in 60 – 80% organic solvent, preferably acetonitrile. (d) System suitability requirements were as follows . — ( 1 ) Resolution. — In the laboratory control sample See Figure 2017.03C, the resolution of the Trp peak and the small interference peak after the Trp peak must be >1.25. ( 3 ) USP tailing. — The USP tailing factor can be calculated within the chromatography software and should be <1.59 for Trp and <1.26 for 5-methyl-tryptophan. Excessive tailing indicates the need for a new column to restore expected performance.

Validation Protocol

A complete SLVwas conducted for this method. The SLVwas conducted following the AOAC Stakeholder Panel on Infant Formula and Adult Nutritionals (SPIFAN) guidelines outlined in SMPR 2014.013. The method was evaluated for precision, accuracy, linearity, analytical range, LOQ, ruggedness, and specificity. Fourteen SPIFAN matrixes were used to validate the method performance.

Precision

All 14 fortified SPIFAN matrixes were prepared and analyzed in duplicate on 6 days. New powder reconstitutions were prepared on each day. Powder samples were blanketed with nitrogen and frozen between uses as specified.

Accuracy

Overspike experiments were conducted on the following SPIFAN matrixes: SRM 1849a, ready-to-feed (RTF) sample 00729RF00 and powder lot Nos. 410057652Z, 00859RF00, K16NTAV, 41045651Z, 00795RF, and E10NWZC. For each run, two predetermined target amounts of the SPIFAN matrixes were weighed into separate 50 mL centrifuge tubes for both low and high overspikes. The samples were subsequently spiked with a Sigma Bovine Serum Albumin (BSA) solution at about 50 and 100% of the previously determined total Trp levels, maintaining the total protein addition of sample plus Bovine Serum Albumin spike to approximately 50 mg maximum. The weight of the spike addition was recorded and subsequently mixed with the sample into the enzyme hydrolysis preparation after all reagents were added. The BSA spiking solution was also prepared for analysis in triplicate with each overspike run. The