AOACSPIFANMethods-2017Awards

132 / 258

1406 Jing et al.: Journal of AOAC International Vol. 98, No. 5, 2015
at the primary transition varied on a given day and a given
instrument. Table 12 shows the results.
Daily ratios for carnitine were, overall, more consistent
than those for choline, especially for data collected on the
TQS instrument. For 4 of the 12 days of TQS choline data,
the transition ratio was approximately twice that observed on
the remaining days. There is no obvious explanation for this
shift. Within-day uniformity was generally good on those days
(within an acceptable range), although on 3 of those 4 days the
ratio RSDs were modestly larger. Despite the large shift in the
ion ratio on those days, however, there was no other indication
that data quality was compromised. Overall, within-day
RSD ranges were carnitine/TQD: 1.3–2.7%, carnitine/
TQS: 0.6–3.6%, choline/TQD: 0.7–1.5%, and choline/TQS:
1.1–5.4%. These are consistent with expectations, generally, for
LC/MS/MS methods.
Because of the differences between instrument platforms as
well as day-to-day shifts, it is not possible to assign meaningful
ranges appropriate for all instruments on all days. A better
alternative is to base the expected ratio in the samples to that
determined for the current daily calibration standards, and to
use some multiple of the SD of the ratio determined for the three
standards to set a limit for the ratio expected in the samples.
Thus, if a particular sample’s ratio was 2–3 SD away (the exact
setting is arbitrary) from the typical ratio in the daily calibration
standards, that would be flagged as a possibly invalid result
(caused by a coeluting molecule that happened to have exactly
the same parent and primary daughter ion mass) and investigated
(e.g., diluted and reanalyzed). The AOAC ERP has previously
suggested this type of approach for MS/MS platforms to deliver
the most highly specific methods to the analytical community.
As reasonable as this approach sounds, we have found it very
difficult to put into practice for LC/MS/MS methods. The flags
raised can disrupt laboratory operations, and the subsequent
investigation/retest is often ambiguous as to whether the result
is more accurate, or even statistically different from the flagged
result. Given the excellent performance of AOAC 2014.04
and good agreement among different methods on a variety of
matrixes, we conclude that confirmatory ion transitions are not
necessary for this method.
Conclusions
An SLV of AOAC 2014.04 has been performed, and the
method has met all the requirements for precision, accuracy,
linearity, LOQ, and specificity as outlined in SMPR 2012.010
for carnitine. The method simultaneously determines choline,
and the data show that it has also met the requirements of SMPR
2012.013 for that analyte.
References
(1) Starkey, D.E., Denison, J.E., Seipelt, C.T., & Jacobs, W.A.
(2) Andrieux, P., Kilinc, T., Perrin, C., & Campos-Giménez, E.,
(3) Fu, S., Tao, B., Lai, S., Zhang, J., & Yiping, R.
(4) Phillips, M.P., & Sander L.C
(5) Oates, K., Chen, L., DeBorba, B., Rohrer, J., Mohindra, D.,
Thermo Fisher Scientific Application Note 193.
(accessed August 9, 2015)
132