Guidelines for Single Laboratory Validation (SLV) of Chemical Methods for Metals in Food
Introduction
The application of analytical methods within a regulatory analysis or accredited laboratory framework imposes certain requirements on both the analyst and laboratory. Under ISO-17025, accredited laboratories are expected to demonstrate both “fitness for purpose” of the methods for which they are accredited and competency of their assigned analysts in performance of the methods[1]. The Codex Alimentarius Commission has issued a general guideline for analytical laboratories involved in the import and export testing of foods which contains four principles[2]:
- Such laboratories should demonstrate internal quality control procedures which meet the requirements of the Harmonised Guidelines for Internal Quality Control in Analytical Chemistry[3];
- Such laboratories should be regular participants in appropriate proficiency testing schemes which have been designed and conducted as per the requirements of the International Harmonized Protocol for Proficiency Testing of (Chemical) Analytical Laboratories[4];
- Such laboratories should become accredited for tests routinely performed according to ISO/IEC-17025:1999 General requirements for the competence of calibration and testing laboratories (now ISO/IEC-170251);and
- Such laboratories should use methods which have been validated according to the principles laid down by the Codex Alimentarius Commission whenever such methods are available.
General requirements for validation of analytical methods according to principles laid down by the Codex Alimentarius Commission are provided in the Codex Manual of Procedures, including provision for “single laboratory” validation of analytical methods[5]. However, there remains considerable misunderstanding among analysts as to precisely what is meant and what is required to demonstrate “method validation”. Additional guidance for possible future inclusion in the Manual of Procedures is currently under discussion in the Codex Committee on Methods of Analysis and Sampling[6]. While compliance with Codex Alimentarius Commission standards and guidelines is voluntary for member states, subject to WTO agreements, they do reflect international consensus on issues discussed. These guidelines can therefore be informative for the development of guidance documents to be used within AOAC International for issues such as single laboratory validation of analytical methods for trace elements.
Validation is defined by ISO as ‘Confirmation by examination and provision of objective evidence that the particular requirements for a specified intended use are fulfilled’ [7]. Method validation has been defined as:
“1.The process of establishing the performance characteristics and limitations of a method and the identification of the influences which may change these characteristics and to what extent. Which analytes can it determine in which matrices in the presence of which interferences? Within these conditions what levels of precision and accuracy can be achieved?
- The process of verifying that a method is fit for purpose, i.e. for use for solving a particularanalytical problem.”[8]
In addition, it is been stated in the IUPAC Harmonized Guidelines for Single Laboratory Validation of Methods of Analysis[9] that:
“Strictly speaking, validation should refer to an “analytical system” rather than an “analytical method”, the analytical system comprising a defined method protocol, a defined concentration range for the analyte, and a specified type of test material.”
Method validation can therefore be practically defined as a set of experiments conducted to confirm that an analytical procedure used for a specific test is suitable for its intended purpose on specific instrumentation and within a specific laboratory environment in which the set of experiments have been conducted. A collaborative study is considered to provide a more reliable indicator of method performance when used in other laboratories because it requires testing of the method in multiple laboratories, by different analysts using different reagents, supplies and equipment and working in different laboratory environments. Validation of a method, even through collaborative study, does not, however, provide a guarantee of method performance in any laboratory performing the method. This is where a second term, verification, is introduced. Verification is usually defined as a set of experiments conducted by a different analyst or laboratory on a previously validated method to demonstrate that in their hands, the performance standards established from the original validation are attained. That is, it meets requirements for attributes such as scope (analytes/matrices), analytical range, freedom from interferences, precision and accuracy that have been identified for suitableapplication of the method to the intended use.
In contrast, method development is the series of experiments conducted to develop and optimize a specific analytical method for an analyte or group of analytes. This can involve investigations into detection/extraction of the analyte, stability of the analyte, analytical range, selectivity, ruggedness, etc. It is important to note that method validation experiments will always take place after method development is complete, in other words, validation studies are to confirm method performance parameters which were demonstrated during method development.
Validation should not begin until ruggedness testing has been completed. A ruggedness design should identify steps of the analytical method where small changes are made to determine if they affect method results. A common approach is to vary seven factors simultaneously and measure these changes to determine how they may affect method performance[10]. Once method development and ruggedness experiments are complete, the method cannot be changed during the validation process.
When validating a method for metals in food products, many factors should be considered during the planning phase of the validation experimental design. For example, is the method to be used in a regulatory environment, and if so, does the analyte of interest have a maximum residue limit (MRL) for which it is assessed for compliance? Is the intended purpose of the method to achieve the lowest possible detection limit? Is the method to be used for the determination of a single element in a particular matrix, or multi-element analyses? Can authentic blank matrix be gathered as the test material? For example many elements are naturally present in a test matrix, such as arsenic in shellfish tissue. The inability to obtain authentic blank test material can cause many validation problems when assessing matrix effects, limits of detection/quantitation, etc.
Although food testing programs frequently include testing for a range of elements (predominantly metals), there are actually few formally established MRLs or other action limits for these analytes. The Codex Alimentarius Commission has established limits for arsenic (total), cadmium, lead in a variety of foods, total mercury in mineral waters and salt, methylmercury in fish and tin in canned goods, as well as for a number of radionuclides in infant and other foods[11]. Similarly, the European Union has established regulatory limits for cadmium, lead, mercury and tin in a variety of foods[12].Requirements for analytical methods to enforce EU standards for lead, cadmium and mercury in foodstuffs are the subject of another EU regulation[13].Canada has established maximum limits for arsenic, lead and tin in various foods[14] and for mercury in seafood[15].
Table 1: Regulated Toxic Elements of Codex and Various Countries
Organization/Country / Regulated ElementCodex / As, Cd, Pb, Hg, MeHg in a variety of foods
EU Countries / Hg, Cd, Pb Sn in some foods
Canada / Hg in fish, Cd, Pb, Sn in some foods
USA / Hg in fish
Japan / Hg and MeHg in some fish
The aim of this single laboratory validation (SLV) protocol is to provide guidance for the scientist when validating a method for inorganic analytes in food or environmental matrices as “fit-for-purpose” for an element or a group of elements in those products. This document provides definitions of common terminology, procedures to be followed, technical guidelines and recommended approaches, as well as an example of a SLV experimental plan. The protocol addresses any specific requirements that are provided in Codex Alimentarius guidance documents or in regulations or guidelines set by national or regional authorities, so is intended to be generally applicable for a variety or potential users.
Definitions
It is recommended that definitions included in the Codex Alimentarius Commission Manual of Procedures5 should be used, when available, as these have been adopted after extensive international consultation and are taken from authoritative sources, such as ISO, IUPAC and AOAC International. A revised list of definitions currently under consideration by the Codex Committee on Methods of Analysis and Sampling (CCMAS) for inclusion in the Codex Manual of Procedures has also been used as a source for the most current definitions which have acceptance within the international analytical science community6.
Accuracy: Closeness of agreement between a measured quantity value and a true quantity value of the measurand[16]. The Codex Manual of Procedures defines accuracy as “the closeness of agreement between a test result and the accepted reference value.”5 The definition currently under consideration by CCMAS 6 is:
“The closeness of agreement between a test result or measurement result and a reference value.
Notes: The term “accuracy”, when applied to a set of test results or measurement results, involves a combination of random components and a common systematic error or bias component. (Footnote: When applied to a test method, the term accuracy refers to a combination of trueness and precision.) Reference:ISO Standard 3534-2: Vocabulary and Symbols Part 2: Applied Statistics, ISO, Geneva, 2006.”
Analytical function: A function which relates the measured value (Ca) to the instrument reading (X) with the value of the interferants (Ci) remaining constant. This function is expressed by the following regression of the calibration results: Ca = f(X)16.
AnalyticalRange:The range of an analytical procedure is the interval between the upper and lower concentration (amounts) of analyte in the sample (including these concentrations) for which it has been demonstrated that the analytical procedure has a suitable level of precision, accuracy and linearity[17].
Applicability6:“The analytes, matrices, and concentrations for which a method of analysis may be usedsatisfactorily.
Note:In addition to a statement of the range of capability of satisfactory performance for each factor, the statementof applicability (scope) may also include warnings as to known interference by other analytes, orinapplicability to certain matrices and situations.
Reference:Codex Alimentarius Commission, Procedural Manual, 17th edition, 2007.”
Bias6: “The difference between the expectation of the test result or measurement result and the true value.
Note: Bias is the total systematic error as contrasted to random error. There may be one or more systematic error components contributing to bias. A larger systematic difference from the accepted reference value is reflected by a larger bias value.
The bias of a measuring instrument is normally estimated by averaging the error of indication over the appropriate number of repeated measurements. The error of indication is the: “indication of a measuring instrument minus a true value of the corresponding input quantity”. In practice the accepted reference value is substituted for the true value. Expectation is the expected value of a random variable, e.g. assigned value or long term average {ISO 5725- 1}.
Reference: ISO Standard 3534-2: Vocabulary and Symbols Part 2: Applied Statistics, ISO, Geneva, 2006.”
Calibration6: “Operation that, under specified conditions, in a first step, establishes a relation between the values with measurement uncertainties provided by measurement standards and corresponding indications with associated measurement uncertainties and in a second step uses this information to establish a relation for obtaining a measurement result from an indication.
Notes: A calibration may be expressed by a statement, calibration function, calibration diagram, calibration curve, or calibration table. In some cases it may consist of an additive or multiplicative correction of the indication with associated measurement uncertainty.
Calibration should not be confused with adjustment of a measuring system often mistakenly called “self calibration”, nor with verification of calibration. Often the first step alone in the above definition is perceived as being calibration.
Reference: VIM, International vocabulary for basic and general terms in metrology, 3rd edition, 2007”
Calibration function: The functional (not statistical) relationship for the chemical measurement process, relating the expected value of the observed (gross) signal or response variable to the analyte amount16.
Certified Reference Material (CRM): A reference material of whose property values are certified by a technically valid procedure, accompanied by,or traceable to, a certificate or other documentation which is issued by a certifying body17.
From CCMAS discussion document6:
“Reference material accompanied by documentation issued by anauthoritative body and providing one or more specified property values with associated uncertainties andtraceabilities, using valid procedures.
Notes:Documentation is given in the form of a “certificate” (see ISO guide 30:1992).
Procedures for the production and certification of certified reference materials are given, e.g. in ISO Guide34 and ISO Guide 35.In this definition, “uncertainty” covers both measurement uncertainty and uncertainty associated with thevalue of the nominal property, such as for identity and sequence. “ Traceability covers both metrologicaltraceability of a value and traceability of a nominal property value.Specified values of certified reference materials require metrological traceability with associated measurement uncertainty {Accred. Qual. Assur., 2006}. ISO/REMCO has an analogous definition {Accred. Qual. Assur., 2006} but uses the modifiers metrological and metrologically to refer to both quantity and nominal properties.
References:
VIM, International vocabulary for basic and general terms in metrology, 3rd edition, 2007.
New definitions on reference materials, Accreditation and Quality Assurance, 10:576-578, 2006.”
Critical value (LC)6: The value of the net concentration or amount the exceeding of which leads, for a given error probability α, to the decision that the concentration or amount of the analyte in the analyzed material is larger than that in the blank material. It is defined as:
Pr ( >LC | L=0) ≤ α
Where is the estimated value, L is the expectation or true value and LC is the critical value.
Notes:
The critical value Lc is estimated by
LC = t1-ανso,
Where t1-αν is Student's-t, based on ν degrees of freedom for a one-sided confidence interval of 1-α and so is the sample standard deviation. If L is normally distributed with known variance, i.e. ν = ∞ with the default α of 0.05, LC = 1.645so.
A result falling below the LC triggering the decision “not detected” should not be construed as demonstrating analyte absence. Reporting such a result as “zero” or as < LD is not recommended. The estimated value and its uncertainty should always be reported.
References:
ISO Standard 11843: Capability of Detection-1, ISO, Geneva, 1997.
Nomenclature in evaluation of analytical methods, IUPAC, 1995.”
Error6: Measured value minus a reference value.
Note:
The concept of measurement ‘error’ can be used both: when there is a single reference value to refer to,which occurs if a calibration is made by means of a measurement standard with a measured value having anegligible measurement uncertainty or if a conventional value is given, in which case the measurement erroris not known and if a measurand is supposed to be represented by a unique true value or a set ot true valuesof negligible range, in which case the measurement error is not known.
Reference:VIM, International vocabulary for basic and general terms in metrology, 3rd Edition, 2007, ISO, Geneva.”
Fitness for purpose6: Degree to which data produced by a measurement process enables a user to make technically and administratively correct decisions for a stated purpose.
Reference: Eurachem Guide: The fitness for purpose of analytical methods: A laboratory guide to method validation and related topics, 1998.”
HorRat6: The ratio of the reproducibility relative standard deviation to that calculated from the Horwitz equation,
Predicted relative standard deviation (PRSD)R =2C-0.15:
HorRat(R) = RSDR/PRSDR ,
HorRat(r) = RSDr/PRSDR ,
where C is concentration expressed as a mass fraction (both numerator and denominator expressed in the same units).
Notes:
The HorRat is indicative of method performance for a large majority of methods in chemistry.Normal values lie between 0.5 and 2. (To check proper calculation of PRSDR, a C of 10-6 should give aPRSDR of 16%.)
If applied to within-laboratory studies, the normal range of HorRat(r) is 0.3-1.3.For concentrations less than 0.12 mg/kg the predictive relative standard deviation developed by Thompson (The Analyst, 2000), should be used.
Reference:
A simple method for evaluating data from an inter-laboratory study, J AOAC, 81(6):1257-1265, 1998
Recent trends in inter-laboratory precision at ppb and sub-ppb concentrations in relation to fitness forpurpose criteria in proficiency testing, The Analyst, 125:385-386, 2000.”
Intermediate Precision:The precision of an analytical procedure expresses the closeness of agreement between a series of measurements obtained from multiple sampling of the same homogeneous sample under the prescribed conditions. Intermediate precision expresses within-laboratories variations: different days, different analysts, different equipment, etc.17
Limit of Detection (LOD): The lowest concentration of analyte in a sample that can be detected, but not necessarily quantitated under the stated conditions of the test16.
Limit of Detection6: “The true net concentration or amount of the analyte in the material to be analyzed which will lead, with probability (1-β), to the conclusion that the concentration or amount of the analyte in the analyzed material is larger than that in the blank material. It is defined as:
Pr ( ≤LC | L=LD) = β
Where is the estimated value, L is the expectation or true value and LC is the critical value.
Notes: The detection limit LD is estimated by,
LD ≈ 2t1-ανσo [where α = β],
Where t1-αν is Student's-t, based on ν degrees of freedom for a one-sided confidence interval of 1-α and σo is the standard deviation of the true value (expectation). LD = 3.29 σo, when the uncertainty in the mean (expected) value of the blank is neglible, α = β = 0.05 and L is normally distributed with known constant variance. However, LD is not defined simply as a fixed coefficient (e.g. 3, 6, etc.) times the standard deviation of a pure solution background. To do so can be extremely misleading. The correct estimation of LD must take into account degrees of freedom, α and β, andthe distribution of L as influenced by factors such as analyte concentration, matrix effects and interference. This definition provides a basis for taking into account exceptions to simple case that is described, i.e. involving non-normal distributions and heteroscedasticity (e.g. “counting” (Poisson) processes as those used for real time PCR). It is essential to specify the measurement process under consideration, since distributions, σ’s and blanks can be dramatically different for different measurement processes. At the detection limit, a positive identification can be achieved with reasonable and/or previously determined confidence in a defined matrix using a specific analytical method.