Agriculture Reference
In-Depth Information
95% of samples. Thus, a false-negative detection rate of 5% is accepted. With regard
to false-positive detects, the method must be veri
ed using an unspiked (blank)
sample of the same commodity. There is no need to specify a criterion for the numbers
of false-positive detects as long as a second sample analysis is to be undertaken using
a second appropriate method for con
rmation of identity.
Cost-effective implementation requires automatic data processing that requires
little or no intervention from the analyst in processing raw data
files. This is usually
achieved by searching the
files against a library containing a database of com-
pounds and associated information such as retention time, chemical formula,
adducts, isotopic patterns, and so on. Optimization of processing parameters
and thresholds is critical. Mol et al. [8] demonstrated the performance and
limitations of a QSMRM based on GC
MS (single quadrupole) detection in routine
use on a variety of fruit and vegetable samples over a 12-month period. Their results
clearly demonstrated the need for regular maintenance of the GC
-
MS system and
ongoing AQC to check performance. In the same paper, the authors also tested a
QSMRM based on UHPLC
-
ToF-MS on samples that had previously been analyzed
using a quantitative MRM based on SRM by LC
-
MS/MS. The latter experiment
also demonstrated the importance of optimizing the thresholds and tolerances of the
software in order to match mass spectrometric and chromatographic information
from the sample extracts with the information in the library. A further publica-
tion [10] describes the analytical capabilities of liquid chromatography with single-
stage high-resolution mass spectrometry (Orbitrap) with respect
-
to selective
detection and identi
cation of pesticides in 21 different fruit and vegetable samples.
This paper clearly demonstrates that the performance of the method is highly
dependent on the instrumentation and software that are employed. In this paper,
high-resolution mass spectrometry allowed analyte detection based on the exact
mass (
5 ppm) of the major adduct ion and of a second diagnostic ion. Using this
two-ion approach, there were only 36 (0.3%) false-positive results from 11,676
pesticide/commodity combinations. The percentages of false negatives, assessed
from 2730 pesticide/commodity combinations, were 13, 3, and 1% at the 0.01, 0.05,
and 0.2mg/kg concentrations, respectively.
The authors used the protocol for method validation as described in the SANCO
document [16] to determine the SDLs for 130 pesticides. These were found to be
0.01mg/kg for 86 pesticides, 0.05mg/kg for 30 pesticides, and
±
0.2mg/kg for 14
pesticides. This paper demonstrates that even when using a high-resolution mass
spectrometer only 66% of the pesticides could be detected at the default MRL value
of 0.01mg/kg [6]. It was suggested that a relative tolerance on the ion ratio
(intensity of ion relative to higher second ion) of
±
50% would be more applicable
than the
30% as stipulated in the SANCO document. Also, the relative retention
time could be reduced from 2min as stipulated in the pesticide document [16] and
±
±
2.5% as stipulated in the veterinary medicine document [18] to
±
1%. Adopting
these values would improve the identi
cation ability of the method at lower
concentrations and reduce the number of false negatives. However, the authors
do warn that the criteria should be checked using data generated from other types of
instrumentssuchasToF.
