AOACSPIFANMethods-2017Awards

127

Jing et al.: J ournal of AOAC I nternational V ol. 98, N o . 5, 2015  1401

Table 4. LOQ verification in SPIFAN sample 11750017V3 for choline at the SMPR limit of 2 mg/100 g RTF a

Theoretical spike, mg/kg

Spike level, mg/100g RTF

Concn (OH-), mg/kg Measured spike b , mg/kg

Sample name

Recovery, %

Adult powder-1 Adult powder-2 Adult powder-3 Adult powder-4 Adult powder-5 Adult powder-6 Adult powder-7

60.14 59.76 61.49 60.60 60.87 59.53 59.99 60.34 233.5 234.9 233.3 232.5 234.1 235.5 232.2 233.7 1.1

Unspiked

RSD, %

Adult powder-LOQ-1 Adult powder-LOQ-2 Adult powder-LOQ-3 Adult powder-LOQ-4 Adult powder-LOQ-5 Adult powder-LOQ-6 Adult powder-LOQ-7

173.1 174.6 172.9 172.2 173.7 175.2 171.8

178.9 178.6 176.5 179.4 180.7 179.0 179.0

96.8 97.8 97.9 96.0 96.1 97.9 96.0 96.9

1.92 1.94 1.92 1.91 1.93 1.95 1.91

Spiked

RSD, %

0.5

0.9

a  For this table, the concentrations measured are in terms of choline hydroxide. b  The mean unspiked value is subtracted from the individual spiked result.

after simple water dilution, and our total carnitine result agrees exactly with the individual NIST result in reference 4). Table 6 shows that there is generally good agreement across the various methods for determination of choline with the exception of the widely used AOAC 999.14 . All four methods compared are quite independent in terms of methodology: microwave-digestion/LC/MS/MS, ion chromatography, enzyme assisted breakdown with colorimetric detection, and HPLC-electrochemical detection. Although this is a limited data set, the obvious recommendation is to avoid using AOAC 999.14 for the analysis of choline in infant formula or adult nutritional products. The other AOAC First Action methods in the table are more accurate over the full breadth of product matrixes presently on the market. Another LC/MS method (AOAC 2012.17 ; 1) was used to confirm the accuracy of AOAC 2014.04 total carnitine results. Twelve internal products (not the SPIFAN set) were analyzed on 8 days by AOAC 2014.04 and on 3–6 days by the reference method. The differences in the mean results for each product ranged from –3.7% to +4.6% with an average of –0.3% (data not shown), indicating the accuracy of this present method for total carnitine. Tables 7 and 8 show spike recovery data from each of the SPIFAN product matrixes at two different levels for choline and carnitine, including a separate set of experiments to prove the recovery of a primary inherent species, acetylcarnitine. No systematic recovery issues were seen, with overall recoveries close to 100% across all matrixes for choline and carnitine. Acetylcarnitine recoveries appeared to be consistently lower, but still within the SMPR-recommended 90–110%.

actual LOQ might be lower due to the sensitivity of the mass spectrometer. LOQ .—Tables 4 and 5 show the results of spiking a placebo product with amounts of choline and carnitine at the SMPR required LOQ. For choline, a recovery of 96.9% was achieved with excellent precision (0.9% RSD). For carnitine, recovery was 105.9% with a 1.4% RSD. The good recoveries and precision indicate that the method is performing at a level well above the LOQ and the SMPRs are thereby met. Indeed, the inherent choline in this product is below the required 2 mg/100 g (6 mg/100 g powder = 0.66 mg/100 g RTF) and is measured with a 1.1% RSD. Accuracy (trueness) .—SRM 1849a was analyzed by the method on 10 days. The result for free carnitine was 13.1 mg/100 g with an RSD of 2.2%, compared to the certified value of free carnitine of 13.6 mg/100 g, a difference of –3.7%. The method result was within the certified range of 12.2–15.0 mg/100 g. A result of 14.9 mg/100 g (1.6% RSD) was also obtained for total carnitine, but the SRM is not certified for total carnitine. The result for total choline in SRM 1849a was 102.7 mg/100 g as choline ion, compared to the certified value of 109.0, a difference of –5.8% but within the certified range of 98–120 mg/100 g. NIST obtained 103.2 ± 0.7 mg/100 g for total choline and 14.9 ± 0.1 mg/100 g for free carnitine using its LC/MS method (4). These latter values are closer to the results of the present method, although it is unclear exactly how the microwave digestion step included in the NIST method affects free carnitine results (our free carnitine result was determined