55 / 71
Pang et al.:
J
ournal of
AOAC I
nternational
V
ol.
98, N
o.
5, 2015
1451
two teas have totally different matrixes. Owing to the different
matrixes, the system deviation of 0.03 s maximum for pesticide
retention time is normal and will not affect the test results. Here,
we mainly focus on learning from Laboratory 23’s pursuit of
meticulous scientific experimentation.
Laboratory 24: In case of contents of 10 μg/kg peak areas
below the areas of the lowest calibration point; oolong tea:
after dissolution of the residues orange and oily drops, the
solution is light yellow and clear. After membrane filtration
solution clear.
Study Director response: Our review of the collaborative
study result table provided by this laboratory has found that the
detection concentrations for the two compounds triadimefon
and trifloxystrobin in green tea incurred samples from this
laboratory turned out to be 40–50% lower than those from
other laboratories, from which we judge that the determination
results of acetochlor are lower than 10 μg/kg, so no specific
detection results are reported. For this problem, we also found
that similar things happened with Laboratory 21. Based on the
above-mentioned descriptions, Laboratory 24 also discovered
that there are orange and oily drops with oolong tea pesticide
extraction solutions, which became clear after filtration
membrane. We also encountered such a phenomenon, which is
normal and will not affect test results.
Laboratory 25: For oolong tea, had to increase amount of
extraction solution to get a better extraction. For oolong tea, the
high standard (the fifth point) had to be removed to have a good
calibration fit (picoxystrobin, indoxacarb, chlorpyrifos, and
butralin). For oolong tea, used linear 1/x to give more weight
to lower standard.
Study Director response: Laboratory 25 proposed
increasing the amount of extraction solution to get a better
extraction for pesticides in oolong tea, and the Study Director
has not discovered similar problems from other laboratories.
The Study Director considers that pesticide extraction
efficiencies have a lot to do with choice of extraction
solutions and amount as well as extraction equipment used.
As for the homogenizer recommended in this collaborative
study method, we compared the efficiencies of three-time
extraction and two-time extraction. Our findings are that
two-time extraction will suffice for the requirements of
recovery indexes. If Laboratory 25 found that a two-time
extraction cannot suffice for good recoveries, an additional
extraction may also be conducted. For calibration curves of
oolong tea, the high concentration points with deviations
for picoxystrobin, indoxacarb, chlorpyrifos, and butralin are
eliminated from their own calibration curves so as to obtain
better calibration curves, which is feasible. For oolong tea,
using liner 1/x gave more weight to lower stand.
Laboratory 28: No results for butralin due to constant signals
in the given MRM transitions.
Study Director response: Regarding no detection of butralin,
our fault is the major cause for the problem. Because instruments
from Waters are unavailable in our laboratory, the precursor ion
m
/
z
337 of butralin given by a local employee from Waters was
not verified, so using 337 is a mistake. In this collaborative
study, Laboratories 9, 22, 29, and 25 also used instruments
fromWaters, and they found 337 is a mistake through scanning,
and the correct one should be 296. Therefore, an accurate
determination was conducted on butralin. In terms of learning
this lesson, the correct ion information can be obtained and such
mistakes can be discovered by scanning the target pesticides
when encountering similar problems in future experiments. If
it was handled like this, this problem could have been solved
earlier. Having said this, the Study Director is not shirking
responsibility but discussing the problem encountered in order
to find out scientific thinking and solve the problem. In a word,
as the Study Director said before, it is our fault and we caused
the trouble to the laboratory involved.
Conclusions
This collaborative study is one of the priority research items
of AOAC in 2010 and a significant systematic project of three
features as follows:
First, review of 6868 test data and related information from
30 laboratories has found that one laboratory deviated from the
collaborative study operational procedures by not undertaking
the pre-study, leading to a great deal of outliers from its sample
inspection; this laboratory was eliminated. A total of 6638
data from the remaining 29 laboratories are effective results
and were inspected with Grubbs and Dixon test to eliminate
187 outliers, making up 2.8% of the total effective data, of
which 1977 were test data from GC/MS determination, with
65 outliers detected, making up 3.3%; 1704 test data were
from GC/MS/MS determination, with 65 outliers detected,
accounting for 3.8%; and 2957 test data were from LC/MS/MS
determination, with 57 outliers detected, making up 1.9%.
Table 14. Distribution range of RSD
r
, RSD
R
, and HorRat values for incurred samples
Parameters
of method
efficacy
RSD
r
, %
RSD
R
, %
HorRat
Range
<8
8–15
>15
<16
16–25
>25
<0.50 0.50–1
.00 1.01–2.00 >2.00GC/MS (16 laboratories)
Green tea
2(100)
a
0
0
1(50)
1(50)
0
0
2(100)
0
0
GC/MS/MS (14 laboratories)
Green tea
2(100)
0
0
0
2(100)
0
0
2(100)
0
0
LC/MS/MS (24 laboratories)
Green tea
0
3(100)
0
0
3(100)
0
0
3(100)
0(0)
0
Total
4(57)
3(43)
0
2(29)
5(71)
0
0
7(100)
0(0)
0
a
Data in parentheses are the percentages.