30 / 71
504
M
astovska
et al
.
:
J
ournal of
AOAC I
nternational
V
ol
. 98, N
o
. 2, 2015
to the blind duplicates) in all seven test samples was 70.5 µg/kg
(RSD = 16.3%,
n
= 56), which corresponds to mean recovery of
88.1%. These results indicate that the mussel and oyster samples
sent to the study participants had good homogeneity. 1,7-DMP
was also added to all shrimp test samples at 20 µg/kg during the
fortification step conducted by the study participants. The mean
concentration value was 22.1 µg/kg (RSD = 19.0%,
n
= 63),
corresponding to mean recovery of 111%. Statistical results for
1,7-DMP obtained in blind duplicate samples are summarized
together with the other analytes in Tables
2014.08B–D
.
Tables
2014.08B–D
provide statistical results obtained for
the studied analytes at three different concentration levels
in shrimp, mussel, and oyster after elimination of statistical
outliers (highlighted in Tables 3–11).
Eight to 10 valid results were obtained for the majority of
determinations. Mean recoveries of all tested analytes at the
total of five different concentration levels were all in the range
of 70–120%: 83.8–115% in shrimp, 77.3–107% in mussel, and
71.6–94.6% in oyster, except for a slightly lower mean recovery
of 68.6% for BaA fortified at 25 µg/kg in oyster (RSD
r
: 5.84%,
RSD
R
: 21.1%) and lower mean recoveries for Ant and BaP
in oyster at all three fortification levels (50.3–56.5% and
48.2–49.7%, respectively).
The lower mean recoveries of Ant and BaP were linked to
degradation of these analytes in oyster samples stored at –20°C
(
see
the discussion about degradation issues below), which also
resulted in lower reproducibility (RSD
R
values in the range
of 44.5–64.7% for Ant and 40.6–43.5% for BaP). However,
the repeatability was good (RSD
r
of 8.78–9.96% for Ant and
6.43–11.9% for BaP), and the HorRat values were acceptable
(1.56–1.94 for Ant and 1.10–1.45 for BaP).
In all other cases, repeatability, reproducibility and HorRat
values were as follows:
(
1
) Shrimp: RSD
r
1.40–26.9%, RSD
R
5.41–29.4%, HorRat
0.22–1.34;
(
2
) Mussel: RSD
r
2.52–17.1%, RSD
R
4.19–32.5%, HorRat
0.17–1.13; and
(
3
) Oyster: RSD
r
3.12–22.7%, RSD
R
8.41–31.8%, HorRat
0.34–1.39.
Overall, the results of the collaborative study demonstrate
that the method is fit-for-purpose to determine PAHs and their
alkyl homologs in seafood samples.
Degradation Issues
The Study Directors reported problems with lower recoveries
for Ant and BaP in oyster samples stored at –20°C to the SPSC
PAH Working Group and the AOAC Methods Committee on
PAHs when they discovered a significant difference between
results for these two analytes obtained in two different
participating laboratories storing the samples at two different
freezer temperatures of –20 and –70°C (
see
recovery results in
Tables 12 and 13, respectively). Due to the limited availability
and cost of oyster samples, it was decided to continue with
the study and not proceed with preparation of new study test
samples that would be fortified by collaborators on the day of
the analysis as was done for shrimp (
Note
: in the case of shrimp,
overall lower recoveries were obtained for all studied PAHs
depending on the shrimp sample storage conditions).
The laboratory storing samples at –20°C analyzed another
set of oyster samples about 1.5 months after the first set
and confirmed the lower recoveries for Ant and BaP. This
collaborator also made another interesting observation related
to the color difference between extracts obtained in the first
set (all dark green) and the second set (the dark green extract
was produced only for the blank sample), whereas all extracts
of fortified samples were yellow-brown. This observation,
which was later confirmed by additional study participants,
indicates matrix changes caused by the presence of PAHs and
accompanied by selective losses of Ant and BaP. In addition to
these two analytes, BaA also showed lower recoveries (around
70%) in oysters when compared to the rest of the studied PAHs.
Furthermore, a closer examination of the results obtained for
Ant, BaP, and BaA in mussel also show somewhat lower mean
recoveries of these three analytes (around 80%) when compared
to the other analytes. These mean results do not include data
obtained for BaP in mussel test samples by Laboratory No. 6,
which analyzed the test samples about a year later than most
other participants. All Laboratory No. 6 BaP results in mussel
test samples showed very good repeatability within the
duplicates but were eliminated as Grubbs’ test outliers because
they were significantly lower than the results obtained by other
laboratories. No other data obtained by Laboratory No. 6 were
identified as outliers using the Cochran or Grubbs’ tests.
Figure 1 provides structures of Ant, BaA, and BaP showing
that these PAHs contain the same moiety in terms of the linear
3-ring (anthracene) structure. This structural commonality
and similar degradation behavior indicate that they could be
substrates for the same enzyme(s). The significant differences in
recoveries obtained for samples stored at –20°C versus –70°C
(Tables 12 and 13, respectively) represent more supporting
evidence for an enzymatic degradation being the most probable
cause for the lower recoveries observed for these analytes in
oyster (and mussel) test samples. To prevent degradation of
these analytes in seafood matrixes during long-term storage,
the Study Directors recommend storing homogenized samples
at –70°C or lower. Unfortunately, this was not a feasible
requirement for this collaborative study because the majority
of the collaborating laboratories did not have this storage
capability.
Acknowledgments
The Study Directors wish to thank the SPSC PAH Working
Group (chaired by Gina Ylitalo from National Oceanic and
Atmospheric Administration), the AOAC Methods Committee
on PAHs (chaired by Tom Phillips, Maryland Department of
Agriculture, Annapolis, MD), and AOAC INTERNATIONAL
staff for the discussions about the study design, analyte
selection, and stability issues as well as for the overall support
provided to the collaborative study. Special acknowledgment
goes to Jack Cochran (Restek Corp., Bellefonte, PA) for useful
discussions about the PAH GC/MS analysis.
John Schmitz and Jack Jabusch (Covance Laboratories,
Madison, WI) and Lucie Drabova and Jana Pulkrabova (ICT,
Prague, Czech Republic) are acknowledged for technical advice
and support during the preparation and the entire course of the
study.
The Study Directors thank the Nutritional Chemistry and
Food Safety business unit at Covance Laboratories for supplying
and preparing PAH standards and spiking solutions; obtaining