Sample Preparation and Matrix Effects in the Detection of

Sample Preparation and
Matrix Effects in the Detection of
Chemical Residues in Foods
Steven J. Lehotay, HyeYoung Kwon,
and Lucía Geis-Asteggiante
Hyeyoung
Kwon
Lucía GeisAsteggiante
Rural Development Administration
National Academy of
Agricultural Science
Suwon, South Korea
Cátedra de Farmacognosia y Productos
Naturales, DQO; UdelaR
Montevideo, Uruguay
Presentation Outline
I. QuEChERS & Update
II. Matrix Effects (GC and LC-API-MS)
III. Pesticide Experiments & Results
IV. Veterinary Drug Residue Results
V. Conclusions
Multiclass, Multiresidue
Analytical Approach:
 extraction
 clean-up
Quick
Easy
QuEChERS method Cheap
Effective
Rugged
 quantification
 identification
Safe
GC-MS
LC-MS
Slide Devised by Katerina Mastovska
QuEChERS Approach
Tomato
Grape
Spinach Strawberry
1) Shake sample with
solvent and salts
2) Centrifuge for 1 min
Spinach
Tomato
Grape
Strawberry
4) Centrifuge for 1 min
5) Analyze Pesticides
3) Mix a
portion
with a
sorbent
Different QuEChERS Methods
2003
Anastassiades et al.
2005
Lehotay et al.
2007
Anastassiades et al.
Original
AOAC 2007.01
CEN 15662
10-15 g sub sample
10-15 g sub sample
10-15 g sub sample



10-15 mL
1% HOAc in MeCN
10-15 mL MeCN
 shake
 shake
 centrifuge
shake & centrifuge
0.4 g/mL anh.MgSO4
0.1 g/mL NaCl
0.1g/mL Na3Cit2H2O
0.05 g/mL Na2Cit1.5H2O
0.4 g/mL anh.MgSO4
0.1 g/mL NaOAc
 shake

 shake
 shake
0.4 g/mL anh.MgSO4
0.1 g/mL NaCl
150 mg/mL anh.MgSO4
25 mg/mL PSA
10-15 mL MeCN
Option:
+ 50 mg
C18 &
7.5 mg
GCB
 centrifuge
150 mg/mL anh.MgSO4
50 mg/mL PSA

shake & centrifuge
 shake
Option:
 centrifuge
+ 50 mg
C18 &
150 mg/mL anh.MgSO4
2.5-7.5
25 mg/mL PSA
mg GCB

shake & centrifuge
Option: Scale-Up & Conc. in Toluene
Slide Devised by Urairat Koesukwiwat
QuEChERS Update
Steve and Angelo “Interviewed” by Ron Majors
QuEChERS a Sample Preparation
Technique that is “Catching On”:
An Up-to-Date Interview with the
Inventors, LC GC North America,
July, 2010
The QuEChERS Revolution,
LC GC Europe, Sept., 2010
Available from ChromatographyOnline.com
QuEChERS in the Literature
Number
100
90
80
70
60
50
40
30
20
10
0
363 Total
Citations to 1st Paper
257 Total
QuEChERS & d-SPE Pubs
According to
ISI Web of KnowledgeSM
*
Year 2003 2004 2005 2006 2007 2008 2009 2010 2011
*search conducted on May 3, 2011
What’s New with QuEChERS?
•
•
•
•
•
•
•
•
•
More vendors and formats (e.g. DPX)
GCB or ChloroFiltr for chlorophyll reduction
Type of MgSO4: can use 97% purity
Shakers – e.g. Spex and Glas-Col
Automation with robotic autosampler
“Unified” method to undergo AOAC update
d-SPE has a life of its own
More applications (e.g. PAHs)
Veterinary drug residue methods
Vendors of QuEChERS Products
b e k o lu t
Disposable Pipette Extraction (DPX)
Patented in 2003 by William Brewer, University of South Carolina
Comparison of ChloroFiltr (16B) and GCB
Recoveries in Lettuces (QuEChERS w/MeCN)
50 mg/mL 16B
7.5 mg/mL GCB
100%
60%
40%
20%
ac
hl
C
or
hl
or
py
rif
os
Fe
nt
hi
on
C
yp
ro
Th
di
ia
ni
be
l
nd
az
Te
ol
bu
e
co
na
zo
le
H
ep
t
zo
lin
lo
il
Vi
nc
ha
l
on
e
ro
t
an
C
hl
o
Li
nd
Q
ui
nt
oz
en
CB
e
0%
H
Recovery
80%
Unified QuEChERS Method
1 g sample per 1 mL of MeCN w/ 1% HOAc
for fruits and vegetables
add internal standard
per g sample, add 0.4 g anh. MgSO4
+ 0.1 g anh. NaOAc
shake or blend
centrifuge
per mL of the upper layer:
150 mg MgSO4 + 50 mg PSA
+ 50 mg C18 + 7.5 mg GCB
mix and centrifuge
QuEChERS for Grains, Nuts, Doughs
2.5 - 5 g sample + 10 mL water*
+ 10 mL MeCN + internal standards
shake for 1 h
add 4 g MgSO4 + 1 g NaCl
shake vigorously for 1 min
centrifuge for 1 min
*15 mL water for 5 g of rice
1 mL of the upper layer
+ 150 mg PSA + 50 mg C18 + 150 mg MgSO4
mix for 30 s
centrifuge for 1 min
H
C
Q
ui M B
nt ire
o
p, ze x
p n
ci o '-D e
s- ,p D
Ch '- E
D
l
o
H rd DT
ep a
C
C
hl tac ne
hl
or h
or
py lo
py
r
r
L
i
B r if in fo
ro o d s
m s-m an
E n C op e e
do yp r op thy
su er yl l
lfa m e ate
n th
S r
Tr ulf in
O ifl at
xy u r e
Pi
rim P flu ali
n
ip erm or fe
ho e n
s- thr
m i
F e et n
Vi nt hy
C ncl hio l
ou oz n
m ol
ap i n
ho
s
Recovery
6 Data Sets from 3 Types of Flaxseeds
QuEChERS of Milled Flaxseeds
100%
Why not correct for recoveries?
80%
60%
40%
20%
0%
n = 26
Biggest Problem with LC- and GC- MS(/MS)
plus costs to purchase and maintain,
and facility requirements, and downtime,
and need for more expertise due to
greater complexity
It is still a pain!
Experiment to Assess Matrix Effects
• 33 LC- and/or GC- amenable pesticides
• 4 matrices (apple, orange, spinach, and rices)
• 20 different sources of each commodity
• Calibration standards from 10-350 ng/g in each
commodity/source and reagent-only
• Analyte protectants added to GC standards
• Analytical sequences conducted on API-3000
LC-(ESI+)-MS/MS and LP-GC/ToF-MS (10 µL PTV)
• Matrix effects calculated (vs. I.S. and not)
How to Calculate (Estimate) Matrix Effects
GC-TOF/MS of Atrazine in Rice
2.0E+06
Reagent Stds
Peak Area
1.8E+06
1.6E+06
Matrix Stds (n=20)
1.4E+06
Linear (Reagent
Stds)
Linear (Matrix Stds
(n=20))
1.2E+06
1.0E+06
y = 4612x + 29698
R2 = 0.9986
y = 4142x + 4558
R2 = 0.9987
8.0E+05
6.0E+05
4.0E+05
%ME = (4612-4142)/4142 = 11%
2.0E+05
0.0E+00
0
50
100
150
200
250
Conc. (ng/g)
300
350
400
Isotopically-Labeled Internal Standard is Ideal
GC-TOF/MS of Atrazine in Rice
Peak Area vs. Atrazine-d5
1.2
Reagent Stds
1.0
Matrix Stds (n=20)
0.8
Linear y(Reagent
= 0.00264x + 0.00132
Stds)
2
R
= 1.000
Linear (Matrix Stds
(n=20))
y = 0.00258x + 0.00124
R2 = 1.0000
0.6
0.4
0.2
%ME = (264-258)/258 = 2%
0.0
0
50
100
150
200
250
Conc. (ng/g)
300
350
400
Alternative Calculation Method
GC-TOF/MS of Atrazine in Rice (w/o I.S.)
40%
20%
10%
0%
-10%
-20%
White Rices
(R13-20)
ME = 13% ± 10%
-30%
-40%
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
Matrix Effect
30%
Brown Rices (R1-R12)
Sample
Effect of Isotopically-Labeled IS
40%
20%
10%
But the I.S. is not infallible
Brown Rices (R1-R12)
0%
-10%
-20%
White Rices
(R13-20)
ME = 2% ± 3%
-30%
-40%
R1
R2
R3
R4
R5
R6
R7
R8
R9
R10
R11
R12
R13
R14
R15
R16
R17
R18
R19
R20
Matrix Effect
30%
GC-TOF/MS of Atrazine vs. IS in Rice
Sample
Results for LP-GC/ToF (w/o I.S.)
100%
Apple
Orange
Spinach
Rice
80%
Matrix Effects
60%
Matrix-Induced Enhancement
40%
20%
0%
-20%
-40%
Ethoprophos
Atrazine
Linuron
Azoxystrobin
Results for LC-MS/MS (w/o I.S.)
10%
0%
Matrix Effects
-10%
-20%
-30%
Ion Suppression
-40%
-50%
Apple
Orange
Spinach
Rice
-60%
Ethoprophos
Atrazine
Linuron
Azoxystrobin
Results for LC-MS/MS (w/ I.S.)
30%
20%
Matrix Effects
10%
0%
-10%
-20%
normalization to atrazine-d5 helps
more than just for atrazine
-30%
-40%
Apple
Orange
Spinach
Rice
-50%
Ethoprophos
Atrazine
Linuron
Azoxystrobin
Results for LC-MS/MS (w/ I.S.)
100%
80%
%ME: orange > spinach > rice = apple
Matrix Effect
60%
40%
20%
0%
-20%
-40%
-60%
-80%
-100%
2.0
Apple
Orange
Spinach
Rice
Linear (Rice)
Linear4.0
(Orange)
3.0
5.0 6.0 7.0 8.0 9.0
Linear (Spinach)
Retention Time (min)
Linear (Apple)
10.0 11.0 12.0
Results for LP-GC/ToF (w/ I.S.)
200%
175%
Matrix Effect
150%
125%
100%
75%
50%
Apple
Orange
cypermethrin
Spinach
Rice
Linear (Rice)
Linear (Orange)
Linear (Spinach)
Linear (Apple)
azoxystrobin
25%
0%
-25%
-50%
2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5
Retention Time (min)
%ME: orange >> rice > spinach = apple
Analyte Protectants
Protectant
Strongly interact with active sites in GC system (inlet,
column and ion source) to decrease degradation and
adsorption of co-injected analytes.
Sharper peaks, less tailing, more ruggedness, lower LOD
HO
O
O
HO
HO
OH
OH
HO
O
HO
OH
OH
HO
OH
HO
ethylglycerol
1 mg/mL
OH
gulonolactone
0.1 mg/mL
O
OH
HO
sorbitol
0.1 mg/mL
shikimic acid
0.05 mg/mL
Effect of Analyte Protectants
w/ analyte protectants
w/o analyte protectants
Anastassiades, Maštovská, Lehotay, J. Chromatogr. A, 1015, 163-184 (2003)
Combination of Analyte Protectants
for GC Pesticide Residue Analysis
O
HO
OH
ethylglycerol (10 g)
Signal enhancement:
moderate
strong
HO
O
HO
HO
O
OH
gulonolactone (1 g)
OH
HO
OH
OH
OH
sorbitol (1 g)
 retention time 
K. Mastovska, S.J. Lehotay, M. Anastassiades, Anal. Chem., 77, 8129-8137 (2005)
Conclusions of Pesticides Study
• Matrix effects aren’t so bad in QuEChERS with LCand GC- MS(/MS) analyses, but worse in citrus
• In terms of matrix effects, one apple is much like
another, and oranges are alike, too, but apples
aren’t like oranges, they’re like plums, etc.
• Analyte protectants in GC improve results, but
matrix-matching still needed for late-eluters,
especially in citrus.
• Isotopically-labeled internal standards work best
to overcome matrix effects, but not perfectly, and
they even help reduce effects for other analytes.
Comparison of 6 Vet. Drug Methods
1) Mol et al. (Rikilt – The Netherlands)
2) Martos et al. (U. Guelph – ON, Canada)
3) Lehotay et al. (USDA-ARS – Wyndmoor, PA)
4) Leepipatpiboon et al. (Chulalongkorn U., Thailand)
5) Stubbings et al. (FERA – York, UK)
6) Kaufmann et al. (Switzerland)
All methods gave similar qualitative MS/MS screening
capabilities with nearly all 60 of the analytes meeting
identification criteria at ½ “tolerance” level in kidney.
Speed, cost, ease of use and ruggedness become the
differentiating aspects.
60 Vet. Drugs in Beef Kidney
vs. SMZ-d6
200%
180%
160%
140%
120%
100%
Mol
80%
Martos
60%
ARS
40%
Thai
20%
Stubbings
e
lin
Te
tr
ac
yc
lfa
s
Su
es
sa
m
id
co
id
e
s+
Lin
w
th
M
ac
ro
l
Gr
o
Fl
u
or
oq
ui
pr
om
no
lo
n
rin
sp
o
al
o
Ce
ph
+
be
ta
-la
ct
am
ot
or
s
0%
Analysis of Incurred Kidney (2 g)
Flunixin vs. Flunixin-d3 IS in Incurred Kidney
700
Concentration (ng/g)
5 min shake; w/o acid
600
5 min shake; w/ acid
500
30 min shake; w/o acid
30 min shake w/ acid
400
60 min heat; w/o acid
60 min heat; w/ acid
300
200
100
0
1
2
3
4
5
6
7
Kidney Sample
8
9
10
Analysis of Incurred Kidney (2 g)
Sulfamethazine vs. SMZ-d6 IS in Incurred Kidney
Concentration (ng/g)
250
5 min shake; w/o acid
200
5 min shake; w/ acid
30 min shake; w/o acid
150
30 min shake w/ acid
60 min heat; w/o acid
100
60 min heat; w/ acid
50
0
1
2
3
4
5
6
Kidney Sample
7
8
9
10
Analysis of Incurred Kidney (2 g)
Penicillin G vs. PenG-d7 IS in Incurred Kidney
Concentration (ng/g)
60
5 min shake; w/o acid
30 min shake; w/o acid
60 min heat; w/o acid
50
40
5 min shake; w/ acid
30 min shake w/ acid
60 min heat; w/ acid
No need for long shake nor heat,
so fast and cool is fine!
30
Spikes were ok, but incurred
penicillins can’t tolerate acids
20
10
0
1
2
3
4
5
6
7
Kidney Sample
8
9
10
Fast Method for Vet. Drug Residues
2 g tissue in a 50 mL tube
add IS mix (SMZ-d6; flunixin-d3; PenG-d7)
add 10 mL of 4/1 (v/v) MeCN/water
vortex briefly, shake for 5 min
centrifuge for 5 min >3500 rcf
supernatant + 500 mg C18 + 10 mL hexane sat’d
w/MeCN; mix for 30 s, centrifuge for
5 min > 3500 rcf; aspirate hexane to waste
evaporate 5 mL extract to 1 mL final vol.
filter extract with the Mini-UniPrepTM
UHPLC-MS/MS analysis
Streamlined Method Validation
Needs:
• Trueness (Recoveries at ≥3 Levels, n > 5)
• Precision (Repeatability & Reproducibility)
• Ruggedness (Multi-day, Multi-Analyst, etc.)
• Selectivity (Interferences in Blanks?)
• Range (calibration and matrix effects)
• Detection limits (MDL, LOD, LOQ, LOI)
• Qualitative (False Negatives/Positives)
Can We Meet All Needs in 3 Days?
3-Day Validation Experiment
Day 1:
• Analyst 1 in hot Lab, Reagents A, 10 matrix blanks
from different sources, 6 spikes at 3 levels each in
6 matrices + 4 spikes each at same levels in mixed
matrices (1 in glass tubes); 5-point calibration
each in mixed matrix and reagent-only stds;
reagent blk = 0-Std inj’d after high std to check for
carry-over
Days 2 and 3:
• Analysts 2 & 3 in cooler labs repeat using
Reagents B & C with different sources of matrices
Veterinary Drug Residues Conclusions
• The streamlined method has met validation
criteria for most drugs in a 3-day validation for
qualitative identification screening purposes.
• Sample throughput is 60 samples/day by 1
chemist for UHPLC-MS/MS analysis.
• The method is being implemented for routine
monitoring of cattle (so far) by the USDA labs.
• Quantitation is acceptable for ≈75% of the drugs,
but enforcement requires 2nd analysis anyway.
• The new streamlined method still needs a cool
name.
Acknowledgments
Michelangelo
Anastassiades
Katerina
Mastovska
Alan
Lightfield
Urairat (Oil)
Koesukwiwat
Terry Dutko and others at USDA FSIS
US-Israel Binational Agricultural Research
and Development Grant US-4273-09
QuEChERS Baby Picture
This is what happens when two fathers
who hate to do dishes have a baby together
Muchas Gracias!
Efficient Pesticide Residue Analysis
Collect Appropriate Sample
(cut and store in dry ice?)
Drive Mobile Lab to the Field
or Send Sample to the Lab
Homogenize and Subsample – Add QC Spk for Sample Processing
Weigh 15 g Sample into Tube (Place ≈200 g for Cold Storage)
Use QuEChERS to Extract (Add I.S.) including d-SPE or DPX Cleanup
250 μL to Mini Uni-Prep Vial - Add QC Std - Transfer 150 µL to Vial w/
Insert for GC; Add Mobile Phase Diluent for LC and Filter in Vial
Inject in (DSI-)LP-GC/MS(-MS)
(Need for Analyte Protectants?)
Inject in UPLC-MS/MS
(Need for Matrix-Matching?)
Incorporate Data into LIMS for QA/QC Review and Report
Conclusions
 QuEChERS is a well-proven, fast sample
preparation method for hundreds of pesticide
residues in different types of food matrices.
 UPLC-MS/MS can provide 10 min analysis of
hundreds of LC-amenable pesticides.
 LP-GC/MS can also provide 10 min analysis of
hundreds of GC-amenable pesticides.
 Currently, the HUGE sample throughput
limitation is data processing and review!
QuEChERS as a Teenager
 QuEChERS is no longer a baby, born of two
fathers, it is a teenager influenced by friends,
some you can trust and others you can’t.
 The QuEChERS approach is still learning its
potential and limitations in the big world.
 QuEChERS concepts are easy and fast to try in
your application(s) – no big loss if it fails.
 Recovery experiments alone are not enough to
validate methods – use incurred samples,
proficiency testing, and/or interlab trials.
Dispersive-SPE
• Why use an SPE apparatus for “chemical filtration?”
• Dispersive-SPE involves the mixing of the sorbent
with the extract in a mini-centrifuge tube to
retain matrix interferants, but not analytes.
QuEChERS Features and Impact
• A single extract can be prepared in 10 min or a batch of
20 in an hr by a single analyst with ≈$1-3 of disposable
materials per sample and generate <12 mL
nonchlorinated solvent waste.
• Consistently high recoveries (mostly 90-110% with
RSDs < 10%) of a wide range of GC- and LC-amenable
pesticides are achieved from many food matrices.
• Countless labs have implemented QuEChERS
successfully for up to 500 pesticides in food and
increased efficiency (faster, less labor, lower cost, less
waste, saves space, less labware, higher throughput).
• QuEChERS concepts have spread to other applications.
What is QuEChERS?
www.quechers.com
http://en.wikipedia.org/wiki/Quechers
The Quechers method is a streamlined
approach that makes it easier and less
expensive for analytical chemists to examine
pesticide residues in food. The name is a
portmanteau word formed from "Quick, Easy,
Cheap, Effective, Rugged, and Safe."
QuEChERS History
2000-2002
2002 EPRW-Rome
2003 Publication
MgSO4
Dispersive-SPE
Analyte Protectants
Limitations of QuEChERS?
•
•
•
•
•
•
•
•
Too Many Modified Versions
Cereals require a separate protocol
Still Problems with captan, folpet, captafol
Spices and oils give problems
Works best with modern MS systems
Need PTV or solvent exchange for low LOD
Matrix effects in complicated matrices
Even simpler sample prep possible
QuEChERS Update
NOTE: Recoveries were normalized to results without ice.
Syringeless Filters
Mini-UniPrepTM (Whatman)
1) Place unfiltered sample
(max. 0.5 mL) in chamber.
2) Compress filter plunger into sample
chamber. Clean filtrate fills reservoir
bottom up.
3) Place the Mini-UniPrepTM vial in an
autosampler.
aqueous samples: PVDF
(polyvinylidenefluoride) filter
QuEChERS Sample Prep
(1) weigh 15 g homogenized sample into a 50 mL tube
(2) add spiking and I.S. solutions, and vortex for 1 min;
(3) add 15 mL of MeCN with 1% HOAc; shake for 30 s;
(4) add 6 g of anh. MgSO4 and 1.5 g of anh. NaOAc;
(5) shake the tube immediately for 1 min;
(6) centrifuge the tube at 3,250 rcf for 2 min;
(7) transfer 1 mL extract to d-SPE tube containing 150 mg
anh. MgSO4 + 50 mg PSA + 50 mg C-18 + 7.5 mg GCB;
(8) mix for 30 s and centrifuge at 3,250 rcf for 2 min;
(9) transfer 0.5 mL into an autosampler vial;
(10) add 50 L of the QC and analyte protectants mixture
and 50 μL MeCN (to make sample volumes equal those
of the calibration standards), and
(11) conduct LP-GC/MS-MS analysis.