COMPARISION of QUANTITATIVE CHEMICAL SINGLE LABORATORY VALIDATION (SLV) REQUIREMENTS

By Jo Marie Cook, Florida Department of Agriculture and Consumer Services

August 2009

Overview:
I reviewed several different validation guidelines and summarized the requirements of each for validation of quantitative, multiresidue type chemical methods. I noted the minimum
requirements for each. Even more replicates are necessary for validation of qualitative methods. The "Guidelines for SLV of Analytical Methods for Trace-Level Concentrations of Organic
Chemicals",was the most complete and relevant guide. It was developed through a harmonization project between CODEX, ISO, AOAC, FAO, IAEA and IUPAC in 1999. I believe it
remains the most important guide for trace level, multi-residue work. All the other guides are more general in nature. This guide also discusses extension of a method or performance
verification by another laboratory. The only significant difference between this guide and others is that most guides recommend calibration for the range of concentrations expected. The
CODEX and EU guides concentrated on ranges within 2 times the regulatory limit.
Most guides mention the need for SLV's and the decline of collaborative studies due to time and cost.
Despite frequent reference to "Fit for Purpose", the need for SLV's and the number of replicates is not significantly diminished for multiresidue methods. If anything, more detailed SLV's are
important due to the scarcity of collaboratively studied methods. Some performance parameters needed to be measured only once such as selectivity, ruggedness and stability. Verification of
method performance in other laboratories required validation parameters of accuracy and precision, similar to those in collaborative studies. However, most of the guides stressed the need to
measure realistic repeatability (multiple days and sample types) and reproducibility (multiple analysts, instruments, reagents, etc). None of the guides mentioned interlab (3 - 5 lab)
validations, presumably because this additional validation information is not statistically significant until 7 or more laboratories are involved.
For most guides, the number of replicates needed for accuracy and repeatability was three blank matrices plus three fortified levels, 5-6 replicates, repeated on three different days. The AOAC
generic protocol was less stringent than any of the other guides.
Because certified reference materials are seldom available, trueness or accuracy must be determined by spiking multiple varieties of blank samples. Inefficiency of extraction can be the major
source of bias. This makes it more important to evaluate method response with standard additions, comparison to alternate methods, proficiencies, evaluate ruggedness, stability etc. It is
important to determine if the method will extract the analytes of interest in incurred samples in the same manner as demonstrated by fortified blank matrix recovery studies.
Most guides mentioned using Youden pairs to determine ruggedness. Variables which contribute significant imprecision should be corrected by additional method development.

References:
AOAC:
AOAC Guidelines for Single Laboratory Validation of Chemical Method for Dietary Supplements and Botanicals, http://www.aoac.org/dietsupp6/Dietary-Supplement-website/DSHomePage2.html, 2002-12-19
Single Laboratory Validation, Generic Protocol (Chemistry) - Draft 8, AOAC, A. Pohland, [email protected]

COMPARISION of QUANTITATIVE CHEMICAL SINGLE LABORATORY VALIDATION (SLV) REQUIREMENTS

By Jo Marie Cook, Florida Department of Agriculture and Consumer Services
August 2009
Single Laboratory Validation Acceptance Criteria (Chemistry Methods) - Draft 8, AOAC, A. Pohland, [email protected]
Protocols for AOAC Method Validation Programs - Rev 2, 3-17-2008, AOAC-OMB, G. Latimer, [email protected]
FDA-CVM:

FDA Guideline 3 V: Guideline for Approval of a Method of Analysis for Residues, FDA, http://www.fda.gov/cvm/Guidance/1731.htm
FDA Guideline 118, Mass Spectrometry for Confirmation of the Identity of Animal Drug Residues, http://www.fda.gov/cvm/Guidance/guide118.pdf
Second Analyst Validation of Study 419.01 - Multiclass Determination and Confirmation of Antibiotic Residues in Honey using LC-MS_MS, Mayda Lopez, FDA CVM
EU:
Commission Decision of 12 August 2002 implementing Council Directive 96/23/EC concerning the performance of analytical methods and the interpretation of results, Official Journal
of the European Communities,
HARMONIZED SLV: (AOAC, Food and Agriculture Organization of the United Nations, International Atomic Energy Agency, International Union of Pure and Applied Chemistry)
Guidelines for Single-Laboratory Validation of Analytical Methods for Trace-Level concentrations of Organic Chemicals, AOAC/FAO/IAEA/IUPAC, 1999
http://www.iaea.org/trc/pest-qa_val2.htm
ISO / IEC: (International Organization for Standardization & International Electrotechnical Commission)
Guidelines for checking and validating test and calibration methods according to ISO/IEC 17025, A011, 1-9-2008, Ver. 4
IUPAC (International Union of Pure and Applied Chemistry:
Harmonized Guidelines for Single-Laboratory Validation of Methods of Analysis, IUPAC Report, Thompson, Ellison, Wood, 11-4-1999
EURACHEM:
The Fitness for Purpose Analytical Methods, A Laboratory Guide to Method Validation and Relate Topics, EURACHEM Working Group, David Holcombe, 1998, ISBN: 0-948926-120
GENERAL READING:
Principles and Practices of Method Validation, edited by A Fajgelj and A. Ambrus, ISBN 0-85404-783-2, Royal Society of Chemistry 2000, www.rsc.org

COMPARISION of QUANTITATIVE CHEMICAL SINGLE LABORATORY VALIDATION (SLV) REQUIREMENTS

By Jo Marie Cook, Florida Department of Agriculture and Consumer Services
August 2009
Performance Parameters

AOAC

FDA-CVM
Guides 3.4, 145-chemistry, 118

EU

Harmonized
Multiresidue SLV Guide

ISO/IEC 17025
A011

DEVELOPMENT and OPTIMIZATION - to be conducted for new methods and may/or may not be included in SLV's by additional laboratories
Fit for Purpose
Fit for Purpose
Purpose
Test "System"
Selectivity or Specificity

Range of Blank matrices

3-4 MS fragments

Retention time
GC/MS=2%, LC/MS=5%
3-4 MS fragments
Full scan 20% match
SIM = 10% match
MS x MS = 20% match

Range of related compounds

(metabolites, derivatives,,,)

Blank spiked w/ interference

False Positive
False Negative
Quant. Enhancement or
suppression
Calibration Curves

required

EUROCHEM Guide

Fit for Purpose

Method, conc. Range, matrix,
use
Selectivity Index,
matrix interference

Fit for Purpose

IP Points

20

5 representative
commodities
5

Bias due to matrix suspected

to be significant

< 5%

zero
< 10%

Zero + 3 levels times 2 10

LCL, AL, AL*2
Non-linear =
7 levels times 3
Matrix vs standards in
solvent
16
Zero, LCL, AL times 5
2-3 * AL times 3

6 levels
0 - 150%, evenly distributed
Run in duplicate or triplicate
in random order
Test matrix effect w/ std. add.
Plot residuals
Estimate, only a rough guide
3 times SD of matrix blank

3 times SD or average of
blank

Detection Capability (LOQ)

10 time SD or average of
blank

Y
Y
Y
Y
5 levels including zero

At least 3 levels

3 times signal to noise

Calibration curve method

5
20 blanks, 3 times signal to
noise
Decision limit plus 1.64 times
the SD of the reproducibility

1, 2, 3, 4 weeks or more
Store at -20C or less
0, 1, 2, 3, 4 weeks (-20C)

Youden pairs

Youden pairs

Slope of calibration curve

Spike interferences
Other methods

Zero + 6 levels
Range of sample levels
Linear level of instrument and
method

Decision Limit (LD or LOD)

Sensitivity

Stability (light, temperature)

Analytes in standard solution
Incurred residue
or matrix fortified
Optimization (re-extraction,
sample preparation, alternate
solvents, shake time/type, pH,
detectors, cleanup, instrument,
columns)
Ruggedness (minor changes)

IUPAC

Youden pairs

16
Zero, LCL, AL times 5
2-3 * AL times 3

Required

Required

10% RSD
or 2 times LOD
Not recommended. State the
uncertainty at concentration

Zero + 6 levels

10 blanks or low level spikes

Terms not generally accepted.
"minimum detectable value"
Blank + 3SD of lowest
concentration
10 blanks or low level spikes
Blanks + 10 SD
May not be lower than the
lowest quantitative measure.

>5 each
Analyte/commodity
& standard solutions

Youden pairs

COMPARISION of QUANTITATIVE CHEMICAL SINGLE LABORATORY VALIDATION (SLV) REQUIREMENTS

By Jo Marie Cook, Florida Department of Agriculture and Consumer Services
August 2009
Comparison to established
method
VALIDATION - basic measurements common to all levels of validation. Some SLV's may require some or all of the tests above.
2 - 3 or more as needed
5
During development - 20
Representative Sample Types
to represent the range of
composition
Trueness (accuracy or bias)
CVM
Certified Ref. Material
or see Recovery
3 levels (upper, lower,
5 levels including zero
Calibration Curves
middle)
(Dietary Supl = 0, 5-7
points)
Recovery
2
9 - 10
20
18
range of expected
0.5, 1, 2 LOQ times 5
1, 1.5, 2.0 LOD times 6
3,4
Analyst performing validation
concentrations
5 incurred
or 0.5, 1, 1.5 Permitted Level
should not be aware of
0 * 3, 15%, 30%, 130%
times 6
concentration / composition
times 3
times 3 days
If possible
Incurred residues
At least 10
<20%
18
Repeatability (% CV)
At least 2 different
Perform the required recovery
Recovery on 3 different days
sample types
experiments over 3 days
(Dietary Supl = 3
Blinded samples
matrices)
Times 2, At least 2 days
15- 20
18
Within-Lab Reproducibility or
Each condition,
Recovery on 3 different days
Intermediate Precision (%CV)
concentrations that differ
(different reagents, analysts,
by order of magnitude
rooms, instruments, days)
times 5
Measurement Uncertainty
6

Ruggedness (major changes)

Required

10

Required
Determine range

Zero, LCL, AL times 5

2-3 * AL times 3

Required

?
2 levels, min and max

18
Blanks + 3 levels times 6

Required

Duplicates times several days,

range of concentration,
different matrix
Horwitz is incorrect < 120
ppb

10

Representative analytes
and matrices

Zero, LCL, AL times 5

2-3 * AL times 3
Representative analytes
and matrices

Required

Inter lab study

Notes:
1. Experiments may be combined to determine different parameters
2. Multiple analytes can be determined at the same time as long as they do not interfere
3. LOD = minimum required quantitative performance limit of the method
4. Recovery is measured by method of standard additions (100 (conc. 2 -conc 1)/concentration added to #2)
5.Concentration at y intercept plus 2.33 times the SD or the reproducibility of the intercept
or 3 times signal to noise
6. Bias, as measured by recovery, is not a component of uncertainty. Bias (a constant) should be removed by subtraction before calculating MU standard deviations (AOAC Dietary Supplement Guidelines for SLV)
7. See FDA Guidance for Industry #118
8. AOAC general guidelines required fewer replicates and levels as recommended by the dietary supplement guide, which was more consistent with other guides.
9. The range of concentrations recommended for recovery and repeatability was narrow. Other guidelines recommended measurements across broader ranges.
10. per the CODEX guidelines: AL = Action Level or Acceptable Level (regulatory limit) and LCL = lowest calibration level

10
Different analysts, equipment,
days

Includes bias if uncorrected