283 / 363
Martin & Campos-Gim
É
nez:
J
ournal of
AOAC I
nternational
V
ol.
98, N
o.
6, 2015
1699
Run in single-reaction monitoring mode. Monitor the
transitions
m
/
z
220.2 → 90.1 for PA, and
m
/
z
224.2 → 94.1
for the isotope-labeled IS, between 0 and 2.1 min. Set collision
energy at 14 V. The dwell time for each monitored transition
is 0.1 s. The last two values are indicative and need to be
checked and optimized for each instrument used.
(d)
Identification
.—MS detection in the single-reaction
monitoring includes simultaneous detection of molecular
ions corresponding to PA and labeled IS. The selected mass
transitions are
m
/
z
220.2 → 90.1 and
m
/
z
224.2 → 94.1,
respectively.
(e)
Quantitation
.—Calculate for each standard the peak area
ratio between PA and IS. Establish a 5-point calibration curve
(ranging from 0.16 to 2.4 ng on column) by plotting peak area
ratio versus PA concentration. Calculate the linear regression.
It is recommended to use a weighted regression curve (1/x).
Calculate the slope (S) and the intercept (I). Calculate the PA
concentration, w, in (mg/100 g) using the following equation:
= (
) ×
1
×
3
× 100
× ×
2
× 1000
where A= peak area ratio PA/IS in the test solution; I = intercept
of the calibration curve; S = slope of the calibration curve;
V
1
= volume of the of sample extract, in mL (= 50); V
2
=
volume of the filtrate pipetted, in mL (= 1); V
3
= final volume of
the test solution, in mL (= 10 ± 1); m = mass of the test portion,
in g; 100 = conversion to 100 g basis; and 1000 = conversion
from μg to mg.
Collaborative Study
Part 1
Participanting laboratories received two practice samples.
Laboratories set up the method described in this paper.
Participants were asked to analyze each of the two practice
samples in duplicate (two extractions from each reconstituted
sample). Any deviation, such as necessity to substitute reagents,
columns, apparatus, or instruments, was to be recorded
and reported. Reporting to the Study Director was done
electronically using a template. Laboratories were asked to give
all areas obtained (both PA and labeled PA) for the standard
curve as well as for the samples. Concerning the standard
curve, participants were given the choice to either use linear
regression or a weighted linear regression (with 1/x as weight).
This decision was to be mentioned in the informatics template.
Furthermore, different masses used during sample preparation
were to be reported. After review by the Study Director,
results within a range of expected levels were used to identify
the laboratories that had the capability to run the analysis
successfully. The laboratories were thus qualified for the second
part of the study.
Part 2
All qualified laboratories received a second shipment
containing 10 products in blind duplicates (i.e., 20 samples)
for the collaborative study. The products came from a set of
infant formula and adult nutritional products (i.e., SPIFAN kit)
aimed to represent the whole range of commercially available
products. Laboratories were asked to analyze all the samples
(single extraction from each reconstituted sample) on 2 days
(10 samples/day). Each sample was assigned to either day 1
or day 2. Results were transmitted to the Study Director via a
similar electronic template as the one used in part 1.
Statistical Evaluation
After data collection, outliers were detected using
Cochran’s and Grubbs’ tests. Average PA concentrations, SDs
of repeatability (S
r
), and RSDs of repeatability (RSD
r
) were
estimated from the blind duplicates in the collaborative study
samples. The duplicates were assigned to be analyzed on the
same day. SDs of reproducibility (S
R
), RSDs of reproducibility
(RSD
R
), and HorRat (Horwitz ratio) values (RSD
R
/predicted
RSD
R
) were also estimated. Details on statistical analysis can
be found in
Appendix D: Guidelines for Collaborative Study
Procedures to Validate Characteristics of a Method of Analysis
of the
Official Methods of Analysis
of AOAC (9).
Table 1. Results of practice samples for 14 laboratories
≤5% ≤15%
Requirements
(SMPR 2012.009)
Mean,
mg/100 g RSD
r
, %
a
RSD
R
, %
b
HorRat
values
Infant formula powder,
milk-based
4.48
2.1
5.3
0.59
Infant formula powder,
soy-based
5.16
2.5
6.0
0.68
a
RSD
r
is the RSD of repeatability.
b
RSD
R
is the RSD of intermediate reproducibility.
Table 2. Results of collaborative study samples for 14
laboratories
≤5% ≤15%
Requirements
(SMPR 2012.009)
n
a
Mean,
mg/100 g RSD
r
, % RSD
R
, %
HorRat
values
Adult nutritional RTF
high-fat
b
14 2.07
2.9
7.0 0.69
SRM 1849a
14 6.96
2.0
5.1 0.60
Child formula powder
14 5.91
2.8
4.9 0.57
Adult nutritional
powder milk protein-
based
13 2.59
1.9
5.0 0.51
Infant formula powder
soy-based
13 5.04
2.8
4.7 0.53
Infant formula RTF
milk-based
13 0.549
1.5
4.1 0.33
Adult nutritional
powder low-fat
13 8.07
1.6
4.1 0.50
Adult nutritional RTF
high protein
13 1.57
1.7
5.5 0.52
Infant elemental
powder
14 6.65
3.3
5.4 0.63
Infant formula powder
part hydrolyzed
soy-based
14 3.85
1.3
5.3 0.57
a
n
= Number of laboratories (after removal of outliers).
b
RTF = Ready-to-feed.
Candidates for 2016 Method of the Year
282